the safety of carotid artery surgery and stenting doig - thesis...the safety of carotid artery...
TRANSCRIPT
The Safety of Carotid Artery Surgery and
Stenting
Submitted for the degree of MD (Res), UCL 2015
David Doig, MBChB, BScMedSci (Hons), MRCP (UK)
2
1. Personal declaration
I, David Doig, confirm that the work presented in this thesis is my own, and has not
previously been submitted for a higher degree. Where information has been derived
from other sources, I confirm that this has been indicated in the thesis.
Dr David Doig MBChB, BScMedSci (Hons), MRCP (UK)
1.1.1 Funding declaration
ICSS was funded by the UK Medical Research Council (MRC) and managed by the UK
National Institute for Health Research (NIHR) on behalf of the MRC-NIHR partnership.
The views and opinions expressed herein are those of the author and do not
necessarily reflect those of the NIHR Health Services Research programme of the
Department of Health. Additional funding was supplied by grants from the Stroke
Association, Sanofi-Synthelabo and the European Union. This work was undertaken at
University College London, which received a proportion of funding from the UK
Department of Health’s National Institute for Health Research Biomedical Research
Centres funding scheme. The funders of ICSS had no role in trial design, data
collection, analysis, or the writing of this thesis.
3
2. Acknowledgements
I would like to acknowledge the huge contribution of the patients who enrolled in the
International Carotid Stenting Study, without whom none of our research would have
been possible. Their involvement with the trial will benefit countless other patients with
carotid artery stenosis.
This work was carried out at UCL under the supervision of Professor Martin Brown and
Dr Rolf Jäger. I thank them both for the opportunity to carry out this work and for their
advice throughout. Dr Leo Bonati of Universitätsspital Basel and Mr Toby Richards of
UCL provided academic support and critical comments on the work. Dr Stefan Brew
and Dr Matt Adams of UCLH NHS Trust provided training in radiological image
analysis. I am grateful to Dr Mark White of UCLH NHS Trust for technical support in
maintaining an imaging database for ICSS analyses. Ben Hobson at UCL collaborated
on several analyses and patiently evaluated hundreds of sets of vascular imaging.
Other ICSS investigators are listed in Appendix II and collaborated on publications
based on this work.
Roland Featherstone is the trial manager for ICSS and assisted with data cleaning and
preparation for analysis. Adrienne Wallis entered trial data. I am grateful to Joanna
Dobson and Liz Turner from the London School of Hygiene and Tropical Medicine who
provided statistical support and advice and helped to produce figures for this thesis.
Helen Tindall organized us all and provided encouragement. Fiona Kennedy gave a
second opinion on analysis results and kept us motivated.
4
3. Personal contribution
The work presented in this thesis is my own, and is based on data from the
International Carotid Stenting Study, a multicentre international trial. A list of trial
investigators and trial centres is given in Chapter 24 (Appendix III). As a UCL Research
Associate I worked in the central trial office from December 2009 to December 2012
cleaning data and designing and performing the analyses set out in Chapters 15, 18
and 19. The analyses in Chapters 14, 16 and 17 were carried out with the practical
assistance and advice of statisticians at the London School of Hygiene and Tropical
Medicine who also helped prepare the forest plots of results as acknowledged in
Chapter 2. While clinical follow-up in ICSS continued I followed up our local trial
patients in person – taking clinical measurements, arranging further investigation and
adjusting their secondary prevention therapy as necessary.
A list of publications and presentations derived from this work is given in Chapters 10-
11.
Where tables or figures appear from other publications, appropriate copyright
permissions have been granted.
5
4. Abstract
4.1.1 Introduction
Carotid endarterectomy (CEA) remains the standard of care for patients with greater
than 50% symptomatic carotid artery stenosis to prevent recurrent stroke. The
International Carotid Stenting Study (ICSS) compared the safety and efficacy of carotid
angioplasty and stenting (CAS), with CEA. This thesis aims to determine which
baseline patient characteristics, processes of care or vascular anatomical variants are
associated with risk of stroke, myocardial infarction (MI) or death within 30 days of
either procedure, and whether existing models to predict the risk of these complications
performed well in ICSS patients.
4.1.2 Methods
ICSS was a multicentre international randomized controlled open trial. Medically stable
patients with >50% symptomatic carotid artery stenosis due to atherosclerosis were
randomized to either CAS or CEA. Procedures were carried out by accredited
operators according to trial protocol.
4.1.3 Results
Women in ICSS experienced a higher risk of stroke, (MI) or death within 30 days of
CEA, although sex was not an independent risk factor for this endpoint. They also
experienced a significantly higher risk of haematoma and cranial nerve palsy after
surgery. The risk of surgery was not significantly influenced by type of anaesthesia,
surgical reconstruction or shunt use.
The risk of stroke, MI or death following CAS rose with increasing age and was higher
with open-cell design stents. The risk of CAS was not influenced by vascular
anatomical variation seen on pre-stenting catheter angiography. Cerebral protection
devices (CPDs), designed to prevent debris causing stroke, did not protect against
periprocedural stroke.
Existing risk scores for CEA and CAS tested in ICSS had insufficient power to
discriminate between low- and high-risk groups.
6
4.1.4 Conclusions
New methods of predicting the risk of carotid revascularization are needed to allow
patients and clinicians to make informed choices about treatment for symptomatic
carotid artery stenosis.
7
5. Table of contents
1. Personal declaration .............................................................................................. 2
2. Acknowledgements ................................................................................................ 3
3. Personal contribution ............................................................................................. 4
4. Abstract ................................................................................................................. 5
5. Table of contents ................................................................................................... 7
6. Table of figures .................................................................................................... 11
7. Table of tables ..................................................................................................... 12
8. Abbreviations ....................................................................................................... 14
9. MeSH Keywords .................................................................................................. 16
10. List of original publications derived from this work ............................................ 17
11. List of original presentations derived from this work ......................................... 19
12. Introduction and literature review ..................................................................... 21
Stroke and vascular disease ..................................................................... 21 12.1
Investigating carotid artery disease ........................................................... 23 12.2
Carotid endarterectomy ............................................................................. 26 12.3
Carotid angioplasty and stenting ............................................................... 30 12.4
From trial data to individual patients: predicting procedural risk ................. 42 12.5
The current status of CAS & CEA for symptomatic patients ...................... 47 12.6
This thesis ................................................................................................. 47 12.7
13. Methods – ICSS and the ICSS-MRI Substudy .................................................. 49
The International Carotid Stenting Study ................................................... 49 13.1
The ICSS-MRI Substudy ........................................................................... 54 13.2
CONSORT flow diagram ........................................................................... 55 13.3
Trial funding .............................................................................................. 55 13.4
14. Risk factors for stroke, myocardial infarction or death following carotid
endarterectomy ........................................................................................................... 57
Introduction ............................................................................................... 57 14.1
8
Methods .................................................................................................... 57 14.2
Results ...................................................................................................... 59 14.3
Discussion and conclusion ........................................................................ 63 14.4
15. Validation of existing risk scores for carotid endarterectomy ............................ 67
Introduction ............................................................................................... 67 15.1
Methods .................................................................................................... 68 15.2
Results ...................................................................................................... 73 15.3
Discussion and conclusion ........................................................................ 82 15.4
16. Risk factors for, and impact of, cranial nerve palsy and haematoma following
carotid endarterectomy ............................................................................................... 85
Introduction ............................................................................................... 85 16.1
Methods .................................................................................................... 86 16.2
Results – cranial nerve palsy .................................................................... 87 16.3
Results – haematoma ............................................................................... 95 16.4
Results – impact of adding CNP to ICSS trial primary outcomes ............. 109 16.5
Results – influence of sex on combined outcomes .................................. 111 16.6
Discussion and conclusion ...................................................................... 113 16.7
17. Risk factors for stroke, myocardial infarction or death following carotid stenting
116
Introduction ............................................................................................. 116 17.1
Methods .................................................................................................. 116 17.2
Results .................................................................................................... 118 17.3
Discussion and conclusion ...................................................................... 127 17.4
18. Vascular anatomical factors influencing the risk of new ischaemic brain lesions
following carotid artery stenting ................................................................................. 132
Introduction ............................................................................................. 132 18.1
Methods .................................................................................................. 132 18.2
Results .................................................................................................... 137 18.3
Discussion and conclusions .................................................................... 148 18.4
19. Validation of an existing risk score for carotid stenting ................................... 151
Introduction ............................................................................................. 151 19.1
9
Methods .................................................................................................. 152 19.2
Results .................................................................................................... 155 19.3
Discussion and conclusions .................................................................... 158 19.4
20. Discussion, conclusions and future directions ................................................ 161
Discussion of key findings ....................................................................... 161 20.1
Conclusions ............................................................................................ 162 20.2
Future directions – who should have what treatment? ............................. 163 20.3
21. Bibliography ................................................................................................... 167
22. Appendix I – ICSS post-treatment care questionnaire .................................... 199
23. Appendix II – ICSS cranial nerve palsy questionnaire .................................... 200
24. Appendix III – List of ICSS centres and investigators ..................................... 202
25. Appendix IV – ICSS trial protocol ................................................................... 207
Protocol summary ................................................................................... 207 25.1
Background ............................................................................................. 208 25.2
Trial design ............................................................................................. 213 25.3
Principal research questions to be addressed ......................................... 222 25.4
Learning curve ........................................................................................ 223 25.5
Effect of changes in technology during the course of the study ............... 223 25.6
Health service research issues ................................................................ 224 25.7
Stenosis after treatment .......................................................................... 225 25.8
Crossovers .............................................................................................. 225 25.9
Data analysis .......................................................................................... 226 25.10
Publication .............................................................................................. 226 25.11
Ethical Committee approval .................................................................... 226 25.12
Data Monitoring Committee ..................................................................... 226 25.13
Steering Committee ................................................................................ 227 25.14
Trial organisation..................................................................................... 227 25.15
26. Appendix V – ICSS-MRI Substudy protocol .................................................... 229
Summary ................................................................................................ 229 26.1
Methods .................................................................................................. 230 26.2
10
Background and aim ............................................................................... 230 26.3
Objectives ............................................................................................... 231 26.4
Study design ........................................................................................... 231 26.5
11
6. Table of figures
Figure 1. Magnetic resonance angiogram of internal carotid artery stenosis ............ 24
Figure 2. The effect of surgery on the absolute risk reduction in NASCET and
ECST at 3, 5 and 8 years’ follow-up stratified by the degree of carotid stenosis at
baseline for a) any stroke or operative death, b) ipsilateral ischaemic or operative
stroke and operative death and c) disabling or fatal ipsilateral or operative stroke and
operative death ........................................................................................................... 28
Figure 3. 3D reconstruction of CT angioram of head and neck of patient illustrating
a carotid stent in situ in the left internal carotid artery (arrowed) ................................. 31
Figure 4. Three cerebral protection devices used in the International Carotid Stenting
Study – Anioguard (top), NeuroShield (middle) and FilterWire EX (bottom) ................ 33
Figure 5. Surgical technique for eversion carotid endarterectomy ............................ 44
Figure 6. ICSS CONSORT trial diagram .................................................................. 56
Figure 7. Univariable predictors of the risk of stroke, myocardial infarction or death
in 821 patients undergoing CEA in whom the procedure was initiated ........................ 61
Figure 8. Receiver operator curve for application of the Rothwell (1999) risk score
to 821 patients undergoing CEA in ICSS .................................................................... 75
Figure 9. Receiver operator curve for application of the Tu (2003) risk score to 821
patients undergoing CEA in ICSS ............................................................................... 78
Figure 10. Receiver operator curve for application of the Kuhan (2001) risk score to
821 patients undergoing CEA in ICSS ........................................................................ 81
Figure 11. Univariable predictors of the risk of cranial nerve palsy within 30 days of
endarterectomy in 821 ICSS per-protocol participants in whom the procedure was
initiated 90
Figure 12. Univariable predictors of the risk of stroke, myocardial infarction or death
in 828 patients undergoing CAS in whom the procedure was initiated ...................... 122
Figure 13. A scanning electron micrograph of closed-cell stent design before (a) and
after (b) expansion, and an open-cell stent before (c) and after (d) expansion. ......... 128
Figure 14. Determination of smooth (D, E), irregular (C) or ulcerated (A, B)
atheromatous plaque at the carotid bifurcation ......................................................... 135
Figure 15. Measuring the length of maximal stenosis in three vascular anatomical
profiles 136
Figure 16. Measuring the total length of stenosis in three vascular anatomical
profiles 153
Figure 17. Receiver-operator curve for the application of the Gröschel (2008) risk
score in 115 patients undergoing CAS in the ICSS-MRI Substudy ............................ 157
12
7. Table of tables
Table 1. Results of follow-up in the early randomized controlled trials of CA(S) vs
CEA 36
Table 2. Results of the recent large randomized controlled trials of CAS vs CEA in
symptomatic patients .................................................................................................. 40
Table 3. Allocation of risk points in the Rothwell risk score ..................................... 70
Table 4. Allocation of risk points in the Tu risk score ............................................... 71
Table 5. Allocation of risk points in the Kuhan risk score ......................................... 72
Table 6. Results of application of the Rothwell (1999) risk score to 821 patients
undergoing CEA in ICSS ............................................................................................ 74
Table 7. Results of application of the Tu (2003) risk score to 821 patients undergoing
CEA in ICSS ............................................................................................................... 77
Table 8. Results of application of the Kuhan (2001) risk score to 821 patients
undergoing CEA in ICSS ............................................................................................ 80
Table 9. Summary of cranial nerve palsies following endarterectomy in ICSS ........ 88
Table 10. Independent predictors of risk of cranial nerve palsy within 30 days of
endarterectomy in 805* ICSS per-protocol participants in whom the procedure was
initiated 94
Table 11. Univariable predictors of risk of haematoma within 30 days of
endarterectomy in 821 ICSS per-protocol participants in whom the procedure was
initiated 96
Table 12. Independent predictors of the risk of haematoma within 30 days of
endarterectomy in 639 ICSS per-protocol participants in whom the procedure was
initiated 108
Table 13. Composite outcome events within 30 days of CEA versus CAS in ICSS
110
Table 14. Results of univariable binomial regression analyses of female vs male in
821 per-protocol CEA patients in whom the allocated procedure was initiated .......... 112
Table 15. Reported immediate cause or timing of event in patients undergoing CAS
in ICSS 120
Table 16. Independent predictors of risk of stroke, MI or death within 30 days of
CAS in 748 ICSS per-protocol participants in whom the procedure was initiated and for
whom complete predictor data are available. Results obtained from multivariable
binomial regression. .................................................................................................. 125
Table 17. Demographic and procedural characteristics of 115 patients allocated to
CAS in the ICSS-MRI Substudy for whom adequate pre-procedure catheter
angiography was available ........................................................................................ 138
13
Table 18. Vascular anatomical characteristics of 115 patients allocated to CAS in
the ICSS-MRI Substudy for whom adequate pre-procedure catheter angiography was
available 140
Table 19. Univariable demographic and procedural predictors of new DWI-MRI
lesions following CAS in 115 patients in the ICSS-MRI Substudy for whom adequate
pre-procedure catheter angiography was available ................................................... 142
Table 20. Univariable vascular anatomical predictors of new DWI-MRI lesions
following CAS in 115 patients in the ICSS-MRI Substudy for whom adequate pre-
procedure catheter angiography was available ......................................................... 145
Table 21. Independent predictors of new DWI-MRI lesions following CAS in 115
patients in the ICSS-MRI Substudy for whom adequate pre-procedure catheter
angiography was available ........................................................................................ 147
Table 22. Allocation of risk points in the Gröschel (2008) risk score .................... 154
Table 23. Results of application of the Gröschel (2008) risk score to 115 patients
undergoing CAS in the ICSS-MRI Substudy ............................................................. 156
14
8. Abbreviations
ACAS the Asymptomatic Carotid Atherosclerosis Study
ACST(-2) the Asymptomatic Carotid Surgery Trial (2)
ARWMC age-related white matter change score
BACASS the Basel Carotid Artery Stenting Study
(s)(d)BP (systolic) (diastolic) blood pressure
CAS carotid artery stenting / carotid angioplasty and stenting
CAVATAS the Carotid and Vertebral Artery Transluminal Angioplasty Study
CCA common carotid artery
CE symbol / mark declaring manufacturer’s product or device meets the
requirement of applicable European Community directives
CEA carotid endarterectomy
(95%) CI (95%) confidence interval
CONSORT Consolidated Standard of Reporting Trials
CREST the Carotid Revascularization Endarterectomy vs Stenting Trial
CT(A) computed tomography (angiogram)
DICOM Digital Imaging and Communications in Medicine
DMC data monitoring committee
DSA digital subtraction angiogram
DWI(-MR) diffusion-weighted imaging (magnetic resonance scan)
ECA external carotid artery
ECG electrocardiogram
ECST(-2) the European Carotid Surgery Trial (2)
EDV end-diastolic velocity (of blood flow)
EU European Union
EVA-3S the Endarterectomy Versus Angioplasty in patients with Symptomatic Severe
Carotid Stenosis trial
FBC full blood count
FLAIR fluid-attenuated inversion recovery
HR hazard ratio
ICA internal carotid artery
15
ICH intracranial haemorrhage
ICSS the International Carotid Stenting Study
ICU intensive care unit
ISRCTN International Standard Randomized Controlled Trial Number
ITT intention-to-treat
MeSH Medical Subject Headings
MI myocardial infarction
mmHg millimetre of mercury
(CE)MR(A) (contrast-enhanced) magnetic resonance (angiogram)
MRC the UK Medical Research Council
MREC Multicentre Research Ethics Committee
mRS modified Rankin Score
N/A not applicable
NASCET the North American Symptomatic Carotid Endarterectomy Trial
OR odds ratio
PP per protocol
PSV peak systolic velocity (of blood flow)
PTA percutaneous transluminal angioplasty
QALY quality-adjusted life year
RR risk ratio
SAPPHIRE the Stenting and Angioplasty with Protection in Patients at High Risk for
Endarterectomy trial
SPACE(-2) the Stent-Protected Angioplasty versus Carotid Endarterectomy trial (2)
TESCAS-C the Treatment of Carotid Atherosclerotic Stenosis in China trial
TIA transient ischaemic attack
UK United Kingdom of Great Britain and Northern Ireland
US(A) United States (of America)
VA the Veterans’ Administration Symptomatic Trial
WHO the World Health Organization
XA catheter / intra-arterial angiogram
16
9. MeSH Keywords
Carotid artery stenosis
Carotid atherosclerosis
Carotid artery ulcerating plaque
Carotid endarterectomy
Stents
Brain ischemia
Stroke
Myocardial infarction
Cranial nerves
Haematoma
17
10. List of original publications derived from this work
10.1.1 Chapter 12 (Introduction and literature review)
Doig D, Brown M M. “Carotid stenting versus endarterectomy”. Annu Rev Med
2012;63:259-276 – work adapted for presentation in this thesis with permission from
the Annual Review of Medicine, Volume 63 © 2012 by Annual Reviews,
http://annualreviews.org
Doig D, Brown M M. “Acute stroke care and management of carotid artery stenosis”.
Pathy’s Principles and Practice of Geriatric Medicine. John Wiley & Sons (2012). pp
655-674
Doig D, Brown M M. “Acute stroke care and management of carotid artery stenosis”.
Cardiovascular disease and health in the older patient. John Wiley and Sons (2012), pp
261-298
10.1.2 Chapter 14 (Risk factors for stroke, myocardial infarction or death
following carotid endarterectomy)
“Risk factors for stroke, myocardial infarction, or death following carotid
endarterectomy: results from the International Carotid Stenting Study” accepted for
publication
10.1.3 Chapter 16 (Risk factors for, and impact of, cranial nerve palsy and
haematoma following carotid endarterectomy)
Doig D, Turner EL, Dobson J, Featherstone RL, de Borst GJ, Brown MM, Richards T.
“Incidence, impact and predictors of cranial nerve palsy and haematoma following
carotid endarterectomy in the International Carotid Stenting Study” Eur J Vasc
Endovasc Surg 2014;48:498-504
10.1.4 Chapter 13 (Risk factors for stroke, myocardial infarction or death
following carotid stenting)
“Predictors of stroke, myocardial infarction, or death within 30 days of carotid artery
stenting: results from the International Carotid Stenting Study” accepted for publication
18
10.1.5 Chapter 18 (Vascular anatomical factors influencing the risk of new
ischaemic brain lesions on diffusion-weighted MRI following carotid
stenting)
“Carotid anatomy does not predict the risk of new ischaemic brain lesions on diffusion-
weighted imaging after carotid artery stenting in the ICSS-MRI Substudy” accepted for
publication
19
11. List of original presentations derived from this work
11.1.1 2013
Featherstone, R. L., Kennedy, F., Doig, D., Dobson, J., & Brown, M. M. “Proportion of
patients treated by carotid endarterectomy or stenting who might have had as good an
outcome with optimised medical treatment: a modelling study using the Carotid Artery
Risk Score”. Cerebrovasc Dis 2013;35:63-63
11.1.2 2012
Doig, D., Featherstone, R. L., Richards, T., Brown, M. M. “Performance of risk
prediction scores for carotid endarterectomy”. Presented at UK Vascular Society AGM,
Manchester, 2012.
Kennedy, F. L., Featherstone, R. L., Doig, D., & Brown, M. M. “Compliance and
durability of blood pressure control for secondary stroke prevention in the International
Carotid Stenting Study”. Cerebrovascular Diseases, 33 (Suppl 2), 1-2. Karger.
doi:10.1159/000339538
Brown, M. M., Dobson, J., Doig, D., Featherstone, R. L., & Turner, E. L.
(2012). “Primary analysis of the International Carotid Stenting Study: a randomised
comparison of the effectiveness of carotid stenting and endarterectomy in preventing
long-term stroke in patients with symptomatic carotid stenosis”. Cerebrovascular
Diseases, 33 (Suppl 2), 1-2. doi:10.1159/000339538
Doig, D. A., Featherstone, R. L., Turner, E. L., Dobson, J., Richards, T., & Brown, M.
M. (2012). “The influence of surgical technique and patient characteristics on the
outcome of carotid endarterectomy (CEA): data from the International Carotid Stenting
Study (ICSS). Cerebrovascular Diseases, 33 (Suppl 2), 1-2. Karger.
doi:10.1159/000339538
Doig, D. A., van der Worp, H. B., Featherstone, R. L., & Brown, M. M. “Timing of
periprocedural major events associated with revascularisation in the International
Carotid Stenting Study (ICSS)”. Cerebrovascular Diseases, 33 (Suppl 2), 1-2. Karger.
doi:10.1159/000339538
20
11.1.3 2011
Doig, D. A., Brew, S., Richards, T., Featherstone, R. L., & Brown, M. M. (2011). “Does
intensive peri-procedural management prevent complications of carotid
revascularization?”. International Journal of Stroke, 6 (s2), 1-65. Wiley.
doi:10.1111/j.1747-4949.2011.00684.x
Featherstone, R. L., Doig, D., Altinbas, A., van der Worp, H. B., & Brown, M. M.
(n.d.). “Impact of 30-day outcomes in the International Carotid Stenting Study (ICSS)
on disability at one year”. International Journal of Stroke, 6 (s2), 1-65. Wiley.
doi:10.1111/j.1747-4949.2011.00684.x
Doig, D. A., Dobson, J., Featherstone, R. L., & Brown, M. M. “Short-term predictors of
stroke after carotid stenting and carotid endarterectomy: data from the International
Carotid Stenting Study (ICSS)”. Cerevrovascular Diseases, Karger.
doi:10.1159/000329448
11.1.4 2010
Doig, D. A., Featherstone, R. L., & Brown, M. M. (2010). Cerebral protection devices do
not protect against peri-procedural stroke: data from the International Carotid Stenting
Study. Cerebrovascular Diseases, 29 ((Suppl 2)), 1-341. Karger.
doi:10.1159/000321266
Doig, D., Featherstone, R. L., & Brown, M. M. “Short-term predictors of risk of stenting
for symptomatic carotid stenosis - data from the International Carotid Stenting Study
(ICSS)”. Journal of Neurology, Neurosurgery and Psychiatry, 82 (3), e1. UK: Highwire
Press British Medical Journal Publishing Group. doi:10.1136/jnnp.2010.235572.11
Featherstone, R. L., Doig, D. A., Altinbas, A., van der Worp, H. B., & Brown, M. M. “The
results of the European Carotid Surgery Trial no longer predict outcome of carotid
endarterectomy: a comparison with results of the International Carotid Stenting
Study”. Cerebrovascular Diseases, 29 (Suppl 2), 1-341. Karger.
doi:10.1159/000321266
21
12. Introduction and literature review
Stroke and vascular disease 12.1
12.1.1 Background
Stroke is one of the leading causes of death and disability throughout the world [1].
Most commonly-defined as “rapidly-developed clinical signs of focal or global
disturbance of cerebral function, lasting more than 24 hours or until earlier death, with
no apparent non-vascular cause” [2], the term “stroke” therefore describes the
presentation of the patient rather than the underlying pathophysiology, and is distinct
from a Transient Ischaemic Attack (TIA) in which the signs and symptoms of focal
cerebral dysfunction resolve within 24 hours. As brain imaging becomes more
accessible, TIA has more recently been defined as “a transient episode of neurological
dysfunction caused by focal brain, spinal cord, or retinal ischaemia without acute
infarction” [3]. Stroke syndromes affect all age groups, becoming increasingly prevalent
in later life. Data from the United States suggest that nearly 800,000 residents in that
country alone suffered a new or recurrent stroke each year [4]. In the United Kingdom
the equivalent figure is estimated to be 152,000 new patients experiencing a stroke
each year [5].
Early treatment and identification of risk factors for recurrent disease is important –
mortality following a stroke may be as high as 41% after 5 years [6]. However, rapid
assessment and organized treatment in a stroke unit, including a multi-disciplinary
approach to rehabilitation, investigation of the underlying cause, attention to secondary
prevention medications and avoidance of complications can improve outcome [7].
There are many causes of stroke or TIA, and identification of the aetiology in each
patient will be guided by the patient’s demographics, the results of clinical examination,
the pattern of infarction or ischaemia in the brain and the results of investigations such
as ECG, cardiac echo, vascular studies and blood tests.
This thesis focuses on patients with ischaemic stroke. In broad categories, ischaemic
stroke may be caused by embolism (from arterial atheroma, from the heart or
paradoxical embolism from the venous system in patients with atrial septal defect),
thrombosis of an intracranial artery, haemodynamic disturbance (particularly in the
setting of carotid or vertebral artery occlusion), or cerebral venous thrombosis [8].
22
12.1.2 Stroke and large-vessel vascular disease
Increasing age is a major risk factor for stroke [9]. A major cause of ischaemic stroke in
the elderly, along with atrial fibrillation, is atherosclerosis affecting the carotid arteries,
the aorta or small penetrating vessels in the brain. The estimate of the proportion of
stroke due to underlying large-vessel disease (largely atheromatous carotid stenosis)
ranges from 13% to 68% in heterogeneous populations [10].
Atheroma is formed wherever there is turbulent blood flow – frequently at sites of
arterial branching such as the carotid arteries, at the origins of the supra-aortic vessels,
in the coronary arteries, the intracranial arteries and the arterial supply to the leg where
there is injury to the vascular endothelium as a result of haemodynamic shear stress
[11]. The response to chronic injury at these sites is infiltration of inflammatory cells
laden with fat (“foam cells”), followed by proliferation of smooth muscle cells. Some
lesions will form a fibrous cap over the top. Reduction in the calibre of the vessel,
stenosis, occurs over time and may give rise to a clinically-audible bruit. Lesions that
are large in volume are more likely to be “unstable” with a high degree of inflammatory
infiltration and a large lipid core [12], leading to rupture of any fibrous cap, exposure of
the circulating blood components to underlying tissue antigens, platelet aggregation
and thrombus formation. Both the thickness of the plaque [13] and irregularity of the
plaque surface [14] have been demonstrated to predict the risk of subsequent stroke in
asymptomatic patients. Progressive occlusion of the artery itself will not usually cause
stroke because of the collateral supply to the brain through the circle of Willis, but if the
circle is incompetent or there is reduction in perfusion due to atherosclerosis of the
other carotid or vertebral arteries then ischaemia may result. However, if thrombus on
the atherosclerotic plaque dislodges following plaque rupture and embolizes distally to
the intracranial arteries then an embolic stroke may result.
Carotid atherosclerosis is relatively common, and can be found in over 2% of European
populations [15]. It is particularly prevalent in patient populations with vascular disease
in the coronary or peripheral arteries, with as many as one in four having co-existent
carotid disease [16] [17]. As expected, therefore, the risk factors for developing carotid
atheroma are similar to those for atheroma in other vessels – older age, smoking, high
blood pressure and high cholesterol [18].
The value of screening patients for carotid disease with a view to primary stroke
prevention is brought into question when the risk of subsequent stroke may be less
than 1% per year in patients using modern secondary prevention medications [19].
23
However, amongst patients who have already had TIA or stroke, those with carotid
atherosclerosis are at highest risk of early recurrence [20]. Targeted secondary
prevention therapy, therefore, consists of aspirin [21] and dipyridamole [22] or
clopidogrel monotherapy [23], antihypertensive medication [24], a statin where
appropriate [25] and control of diabetes. Patients are also encouraged to modify their
lifestyle risk factors for stroke by eating healthily, taking part in moderate-intensity
exercise, smoking cessation and reducing alcohol intake [26].
Investigating carotid artery disease 12.2
Carotid stenosis is difficult to detect clinically. Turbulent flow across a stenosis may
produce a bruit clinically-detectable using a stethoscope, although high degrees of
stenosis may reduce flow to a degree that no bruit is audible. Additionally, less-relevant
external carotid artery (ECA) stenosis may also produce a bruit, as may transmitted
sounds from an ejection systolic heart murmur. As such it is an unreliable sign and
further radiological investigations are required to confidently make the diagnosis.
12.2.1 Catheter angiography
Digital subtraction catheter angiography (DSA or “XA”) is often regarded as the “gold
standard” investigation for the diagnosis of carotid atheroma, and was the primary
method of determining the degree of stenosis in most of the early clinical trials of
carotid surgery versus medical therapy for symptomatic carotid stenosis. A catheter is
advanced from a peripheral artery, usually the femoral, around the aortic arch. The
common carotid is then cannulated and the carotid arteries imaged by means of
radiopaque dye. A “mask” image obtained before the administration of intra-arterial
contrast is then digitally subtracted from the contrast image to remove other radio-
opaque structures such as bone from the final image. However, this invasive technique
carries with it a risk of stroke or TIA of up to 4% in those with even lower-risk
asymptomatic plaques [27] due to dislodgement of either atheroma in the vascular tree
or thrombus on a complicated plaque. Figure 1 illustrates a stenosis of the left internal
carotid artery just above the level of the carotid bifurcation (arrowed) as viewed on
magnetic resonance angiography.
24
Figure 1. Magnetic resonance angiogram of internal carotid artery stenosis
An atheromatous plaque causing stenosis of the internal carotid artery just above the
level of the carotid bifurcation is arrowed
25
12.2.2 Ultrasound
Because of their proximity to the surface of the neck, it is usually possible to examine
the internal and external carotid arteries as they branch from the common carotid artery
using ultrasound. Ultrasound may be used to visualize the internal carotid artery (ICA)
itself, and although features such as plaque thickness can be measured, the degree of
stenosis is usually is most closely correlated to the velocity of blood flowing through the
artery. As the systolic blood pulse flows through a stenosis there is a corresponding
increase in the velocity of the blood which is measured using Doppler-mode
ultrasound. The degree of stenosis is derived from measurement of the peak systolic
velocity (PSV) and end-diastolic velocity (EDV) of blood flow in the common carotid
and internal carotid arteries and the ratio between the two measurements, although
technical limitations to determining the degree of stenosis include acoustic shadowing
from calcification in the plaque and difficulties interpreting velocity measurements in
tortuous vessels or arteries without normal runoff [28].
12.2.3 CT and MR angiography
Computed tomography (CT) angiography is a less-invasive alternative to DSA. Risks
associated with CT angiography include contrast nephropathy and allergic reaction to
contrast media, and the ability to visualize the degree of stenosis may be limited in
lesions with heavy calcification that can cause artefact on the scan image.
Magnetic resonance angiography (MRA) has a number of advantages over CT,
including the avoidance of ionising radiation. However, patients do not tolerate the
investigation as well, mostly because of the confined nature of the scanning tunnel.
Contrast-enhanced MR has been reported to be more accurate investigation than CTA
or DUS for the detection of high-grade (70-99%) stenosis than other non-invasive
methods (>90% sensitivity and specificity) [29], although it is not clear whether
combinations of non-invasive tests are as reliable for detecting carotid stenosis
compared with catheter angiography as the results of different investigations do not
always agree on the degree of stenosis. As MR angiography without contrast (“time of
flight” sequences) can lead to over-estimation of the degree of stenosis [30] contrast-
enhanced MRA is usually preferred. Because of the risk of complications associated
with catheter angiography, non-invasive tests, sometimes in combination, are currently
preferred for the initial investigation of carotid stenosis, with catheter angiogram being
reserved for those cases where doubt over the diagnosis or degree of stenosis
remains.
26
Carotid endarterectomy 12.3
That carotid artery stenosis could be a cause of focal neurological symptoms has long
been recognized. Early forms of surgical intervention were focussed on attempts to
improve blood flow through techniques such as cervical sympathectomy, based on the
assumption that a lack of blood supply to the ipsilateral cerebral hemisphere produced
symptoms (“brain claudication”) similar to critical reduction of blood flow in the coronary
arteries causing symptoms of angina through myocardial ischaemia. The first published
descriptions of carotid endarterectomy – surgical removal of the atheromatous lesion –
were from the surgeons DeBakey in the United States and Eastcott in the United
Kingdom. In 1975 DeBakey reported having operating on a patient in 1953 with
recurrent left hemisphere ischaemic symptoms due to carotid atheroma with good
results [31]. Eastcott in 1954 reported endarterectomy and repair of the carotid artery
on a 66-year-old housewife with similar ischaemic symptoms [32].
Endarterectomy for carotid stenosis was widely adopted before large-scale clinical
trials could provide definitive evidence for its efficacy in the prevention of recurrent
stroke, and by the early 1980s had become the third-most performed operation in the
United States [33]. Surgery for carotid stenosis carries a risk of complications, including
death, stroke, myocardial infarction, wound haematoma, cranial nerve palsy and
pulmonary problems [34], and so to justify performing the procedure these short-term
risks must be outweighed by the long-term benefit of the operation in terms of stroke
prevention.
12.3.1 Trials of carotid endarterectomy versus medical therapy – symptomatic
patients
In the 1980s and 1990s two large clinical trials were performed which provide the
foundation for evidence-based intervention in recently-symptomatic patients since – the
North American Symptomatic Carotid Endarterectomy Trial (NASCET) [35] and the
European Carotid Surgery Trial (ECST) [36].
NASCET grouped patients into those with “low moderate” (<50%), “high moderate” (50-
69%) and “severe” (70-99%) carotid stenosis. Patients were randomized to either CEA
or medical treatment. Interim analysis of 659 patients with 70-99% stenosis revealed a
statistically significant reduction in the risk of any stroke or death over 2 years in the
surgery group (15.8%) as compared to the medical group (32.3%) [37]. The
subsequent observation that the benefit of surgery in terms of stroke prevention was
27
reduced in groups with lower degrees of stenosis (and, indeed, absent in those with
<50% stenosis) led the authors to conclude that “the degree of benefit individual
patients received from carotid endarterectomy was directly proportional to the risk they
faced without surgery” [35], with lesser degrees of stenosis conferring lower risk of
recurrent stroke.
ECST, a trial that enrolled 3204 patients, reached a similar conclusion, but used a
different method of measuring stenosis [36]. This trial showed a significant reduction in
the risk of stroke or surgical death over 5 years in those undergoing CEA vs those on
medical therapy of 21.2% (95% CI 12.9% to 29.4%) in patients with 70-99% stenosis
and 5.7% (95% CI 0% to 11.6%) in those with 50-69% stenosis [38]. Like NASCET,
there was no evidence of a benefit of surgery in patients with less than 50% carotid
stenosis.
Meta-analysis of the results of NASCET, ECST and the smaller 189-patient trial from
the Veterans Affairs Cooperative Studies Program 309 Trialist Group [39], using only
the NASCET method of measuring stenosis [40], clarifies the extent of this risk
reduction in patients with varying degrees of stenosis. Overall, the absolute risk of
ipsilateral stroke was 16.0% lower (p<0.01) in those with more than 70% stenosis
undergoing CEA as compared with those in the medical treatment group, and 4.6%
lower (p=0.04) in patients with 50-69% stenosis undergoing surgery [38]. Patients with
less than 50% stenosis did not benefit. A specific subgroup of patients with greater
than 99% stenosis and vessel collapse due to poor runoff distal to the stenosis also did
not benefit from surgery, perhaps because these lesions progressed to complete
occlusion [41]. The relationship between the reduction in the main outcomes of these
trials with CEA and the degree of stenosis is illustrated in Figure 2.
28
Figure 2. The effect of surgery on the absolute risk reduction in NASCET and
ECST at 3, 5 and 8 years’ follow-up stratified by the degree of carotid stenosis
at baseline for a) any stroke or operative death, b) ipsilateral ischaemic or
operative stroke and operative death and c) disabling or fatal ipsilateral or
operative stroke and operative death
Adapted from [38] and reproduced from [42] with kind permission of Elsevier and of the
Annual Review of Medicine, Volume 63 © 2012 by Annual Reviews
29
There is evidence that some subgroups of patients in whom surgery is recommended
will benefit more than others. The risk/benefit ratio most favours surgery over medical
treatment in men and the elderly. In addition, the overall benefit of surgery diminishes
as the time between symptoms and surgery increases, strongly favouring intervention
within 2 weeks of symptoms in stable patients [43].
The results of these trials, taken together, demonstrated a strong link between
increasing degrees of carotid stenosis and increasing risk of recurrent stroke or TIA on
medical treatment alone. This was matched by a corresponding decrease in risk for
patients with high degrees of stenosis who undergo CEA. The most recent guidelines
from the American Heart Association [44], the Society for Vascular Surgery [45] and
the European Society for Vascular Surgery [46] continue to recommend carotid
endarterectomy for the long-term prevention of stroke in patients with recently-
symptomatic carotid artery stenosis >50% and in selected asymptomatic patients with
60-99% stenosis. However, stroke, myocardial infarction and death are not the only
risks associated with CEA, and both cranial nerve palsy [47] [48] [49] [50] and neck
haematoma [51] [52] are well-recognized “minor” complications which may have a
major impact on the patient but have been less-extensively studied in clinical trials.
These extra hazards of a neck incision have led to the consideration of carotid stenting
as a less-invasive alternative to CEA, and as an alternative in patients with difficult
surgical access owing to previous neck surgery or carotid stenosis secondary to
radiotherapy [53].
12.3.2 Trials of carotid endarterectomy versus medical therapy – asymptomatic
patients
The rate of ipsilateral stroke in patients with asymptomatic carotid disease is much
lower than that of patients with symptomatic stenosis – possibly as low as 0.34% per
year when optimal medical therapy is prescribed [19]. Two large randomized trials
provide much of the data addressing the question of whether prophylactic CEA in
asymptomatic patients prevents long-term stroke.
The North American Asymptomatic Carotid Atherosclerosis Study (ACAS) randomized
1662 patients with ≥60% stenosis to either daily aspirin treatment or daily aspirin plus
carotid endarterectomy. In addition, patients in both groups received “medical risk
factor management” targeting vascular risk factors. The primary comparison was
between the rate of ipsilateral stroke or any perioperative stroke or death. CEA reduced
30
the risk of this outcome from 11.0% to 5.1% (relative risk reduction 53%, 95% CI 22%
to 72%) [54].
The Asymptomatic Carotid Surgery Trial (ACST) randomized 3120 patients with 60-
99% carotid stenosis to either “immediate” endarterectomy or deferral of CEA until a
clinician considered there to be a clear indication for proceeding with surgery. Half of
the patients in the “immediate” CEA group were operated on within one month of
randomization. The combined risk of perioperative and subsequent stroke over a five-
year period was 11.8% in the deferred group and 6.4% in the immediate surgery group
(p<0.01) [55]. Long-term results have since confirmed an extended benefit of CEA
stroke prevention for patients under the age of 75 years [56] up to 10 years after
surgery.
Carotid angioplasty and stenting 12.4
Carotid angioplasty and stenting evolved as a less-invasive alternative to CEA in part
to avoid the hazards of a surgical incision. One of the first descriptions of the use of
endovascular intervention for atheroma in a peripheral artery by balloon angioplasty
was by Dotter in 1964, who recanalized the femoral artery of an elderly patient with
peripheral vascular disease who had critically-reduced perfusion to the foot. The
procedure was a clinical and radiological success [57]. (He would later develop the
technique further by introducing the first stents (“coilspring grafts”) into the femoral or
popliteal arteries of dogs, demonstrating that the stents remained patent after insertion
[58]).
The first descriptions of using this technique to treat atheromatous disease of the
carotid artery in humans emerged in the 1980s, although only small numbers of
patients were included in published case series. Early results were promising, although
the problem of transient neurological and cardiac symptoms due to balloon inflation
and the hazards of carotid artery dissection were recognized [59] [60]. A significant
restenosis rate was also reported [61], although early restenosis after angioplasty often
represents smooth muscle growth rather than enlargement of the atheromatous plaque
and therefore was thought to confer a lower risk of subsequent ischaemic stroke.
However, much like endarterectomy, CAS was widely adopted before good-quality
evidence from large randomized trials was available to support its routine use. By 2003
an estimated 10,000 patients worldwide had been recorded as having undergone CAS
with a reported rate of recanalization over 99% [62]. Figure 3, below, shows a 3D
reconstruction of a carotid stent in situ in the internal carotid artery (arrowed).
31
Figure 3. 3D reconstruction of CT angioram of head and neck of patient
illustrating a carotid stent in situ in the left internal carotid artery (arrowed)
32
One of the major developments in CAS followed the observation that particulate debris
generated during the procedure itself was a cause of periprocedural stroke. “Cerebral
protection devices” (CPDs) were developed to catch this debris before entering the
cerebral circulation, and employed either a mesh-type filter to trap particles or used
proximal and/or distal balloons to obstruct flow. There is some conflicting evidence as
to the efficacy of CPDs in preventing procedural cerebral ischaemia, and while one
clinical trial found poor results in patients who did not receive a CPD [63], the
International Carotid Stenting Study demonstrated a significantly higher rate of new
ischaemic brain lesions after CAS versus CEA in patients recruited in centres with a
policy of routine CPD use (OR 12.20, 95% CI 4.53 to 32.84) compared with CAS
versus CEA in centres using unprotected stenting (OR 2.70, 95% CI 1.12 to 6.24,
interaction p=0.02) [64]. Flow reversal systems, first described by Parodi [65], are a
newer alternative to filter-type CPDs that use a system of inflatable balloons to create
reversal of flow in the ICA into an arteriovenous shunt created by the operator. This
technique has been less extensively studied, but is currently undergoing evaluation
[66]. Figure 4, below, illustrates three cerebral protection devices used in ICSS.
33
Figure 4. Three cerebral protection devices used in the International Carotid
Stenting Study – Anioguard (top), NeuroShield (middle) and FilterWire EX
(bottom)
From [67] with kind permission from Lippincott Williams & Wilkins / Wolters Kluwer
Health
34
Stenting carries hazards because of the endovascular nature of the procedure.
Atheromatous lesions around the level of the carotid bifurcation are likely to contain
friable material [68] which can be dislodged by catheterization or the passage of
endovascular devices across the lesion causing TIA or stroke. Although avoiding an
incision in the neck, there is still a risk of groin haematoma of around 2-3% which may
necessitate repair of the femoral artery used for vascular access [69]. Patients who
may have renal impairment should have their renal function checked prior to
angiography and/or stenting as the procedure involves the administration of iodinated
intravascular contrast.
12.4.1 Trials of CAS vs CEA – symptomatic stenosis
Early randomized trials comparing CAS and CEA yielded mixed results and generated
some concern about the safety of CAS. In Leicester in the UK, Naylor and colleagues
randomized 23 patients (of whom 17 received treatment) to CEA or angioplasty before
the study was stopped owing to a periprocedural stroke risk of 71% (5/7 patients) in the
stenting arm [70]. In contrast to this risk, Brooks and colleagues in Lexington,
Kentucky, recorded only one death out of 51 patients randomized to CEA and one TIA
in 53 patients randomized to CAS in their trial [71]. They noted the shorter length of
hospital stay in CAS patients and concluded that CAS “challenged” CEA as the
preferred treatment for patients with symptomatic carotid stenosis. The BACASS trial in
Basel, Switzerland, reported similar results – none of the 10 patients randomized to
CAS in this study suffered stroke, myocardial infarction or death before the trial was
stopped when the centre decided to participate in the International Carotid Stenting
Study [72]. The results of the WALLSTENT study [73] have only been published in
abstract form, but it is known that after enrolling 219 patients the risk of any stroke or
death within 30 days in the CAS arm was 12.1% vs 4.5% for CEA (p=0.05). The study
was stopped on the grounds of futility.
Two larger multicentre trials more firmly established the basis for offering angioplasty
as an alternative to CEA – the Stenting and Angioplasty with Protection in Patients at
High Risk for Endarterectomy (SAPPHIRE) trial [74] and the Carotid and Vertebral
Artery Transluminal Angioplasty Study (CAVATAS) [75].
CAVATAS included patients with anterior circulation symptoms due to carotid stenosis
as well as those with posterior circulation symptoms due to vertebral stenosis. Patients
were randomized to either angioplasty or endarterectomy. The 30-day rate of stroke
with symptoms persisting more than 7 days or death was almost identical between the
35
two groups [75], but during extended follow-up there was a non-significant increase in
the 8-year incidence of ipsilateral non-perioperative stroke risk to 11.3% in the
endovascular group versus 8.6% in the CEA group [76]. This study of 504 patients was
therefore considered underpowered to detect a difference in treatment effect, and
interpretation of the result was complicated by the introduction of stenting as a routine
addition to angioplasty during the course of the trial.
SAPPHIRE enrolled symptomatic patients considered at “high risk” for CEA – those
with cardiac or respiratory disease, contralateral carotid occlusion, recurrent stenosis,
previous neck surgery or radiotherapy, age over 80 years or contralateral recurrent
laryngeal nerve palsy – who had stenosis of at least 50%. Asymptomatic patients fitting
the above criteria were enrolled if they presented with carotid stenosis greater than
80%. Patients were randomized between CAS and CEA, and CAS patients were
treated using a CPD. At one year, the specified endpoint of death, stroke or MI was
reached by 12.2% of CAS and 20.1% of CEA patients, but confidence intervals for this
difference included equality of the two treatments [74]. At three years follow-up there
was also no difference in the risk of the combined endpoint.
The results of the long-term follow-up in these and one other early trial of CAS vs CEA
is presented in Table 1 below.
36
Table 1. Results of follow-up in the early randomized controlled trials of CA(S) vs CEA
Trial Region N Inclusion
Criteria
Comparison Primary Outcome Follow-up Risk of Primary Outcome
Naylor et al.
[70]
UK 23 >70%
symptomatic
ICA stenosis
Angioplasty
vs
CEA
Death or disabling or non-disabling
stroke
30 days 71% vs 0% (p<0.01)
Brooks et al.
[71]
USA 104 >70%
symptomatic
ICA stenosis
CAS
vs
CEA
Death or cerebral ischemia 2 years 1.9% vs 2.0%
Alberts et al.
(WALLSTENT)
[73]
USA 219 60 – 99%
symptomatic
ICA stenosis
CAS
vs
CEA
Ipsilateral stroke, procedure-related
death, or vascular death within 1 year
1 year 12.1% vs 3.6% (p=0.02)
37
Table 1 continued
Trial Region N Inclusion Criteria Comparison Primary Outcome Follow-up Risk of Primary Outcome
Hoffman et al.
(BACASS) [72]
Switzerland 20 ≥70% symptomatic
ICA stenosis
CAS
vs
CEA
Periprocedural stroke, death or MI 4 years 0.0% vs 10.0%
Ederle et al.
(CAVATAS) [75]
International 504 Carotid stenosis
requiring treatment
(90%
symptomatic)
Angioplasty ±
stenting
vs
CEA
Ipsilateral non-perioperative stroke 8 years 11.3% vs 8.6% (HR 1.22,
95% CI 0.59-2.54) by
intention to treat
Yadav et al.
(SAPPHIRE)
[74]
North
America
334 ≥50% symptomatic
stenosis / ≥80%
asymptomatic ICA
stenosis
CAS
vs
CEA
Death, stroke or MI within 30 days
or death or ipsilateral stroke
between 31 and 1080 days
3 years 24.6% vs 26.9% (absolute
difference -2.3%, 95% CI –
11.8% to 7.0%) by
intention to treat
38
Four large randomized controlled open trials of CAS versus CEA that enrolled
symptomatic patients have been carried out more recently, and all have formally
reported results.
The Stent-Protected Angioplasty versus Carotid Endarterectomy (SPACE) trial was
carried out in Central Europe. Of non-inferiority design, it enrolled 1200 patients with
symptomatic “severe” stenosis (≥50% by NASCET method [40]) before analysis
determined that it has failed to conclusively demonstrate non-inferiority. The trial was
thus stopped early when its initial funding source expired and it was thought unlikely to
reach a conclusive sample size. Analysis of 1183 patients showed a 30-day risk of
death or ipsilateral stroke of 6.84% of the CAS group and 6.34% of the CEA group
(absolute risk difference 0.51%, 90% CI -1.89% to 2.91%) [77]. At 2-year follow-up
there was a similar rate of recurrent ipsilateral stroke in both groups, although it was
noted that recurrent carotid stenosis was more common in the CAS group [78].
The French Endarterectomy versus Angioplasty in Patients with Symptomatic Severe
Carotid Stenosis (EVA-3S) trial was also of non-inferiority design. After randomization
of 527 patients recruitment was stopped on the advice of the safety committee – the
primary combined endpoint of periprocedural stroke or death occurred in 9.6% of the
CAS group (95% CI 6.4% to 14.0%) compared with 3.9% of the CEA group (95% CI
2.0% to 7.2%) [79]. Follow-up was continued out to four years, and showed that the
subsequent long-term risk of recurrent ipsilateral stroke, excluding procedural risk, was
low and “similar in both treatment groups” [80].
The International Carotid Stenting Study (ICSS) was funded mainly by the UK Medical
Research Council, the Stroke Association and the European Union, and it recruited
internationally. It randomized 1713 patients with greater-than 50% carotid stenosis
using the NASCET method of measurement or non-invasive equivalent [81]. The trial
protocol is described in brief in Chapter 13 and presented in full in Appendix IV. In an
intention-to-treat analysis of results up to 120 days after randomization the risk of
stroke, myocardial infarction or death in the CAS group was 8.5% versus 5.2% in the
endarterectomy group (HR 1.92, 95% CI 1.27 to 2.89) [82]. Recruitment finished in
2008 and the results of long-term follow-up have been presented. The long-term risk of
fatal or disabling stroke was no different between the two groups during a median
follow-up duration of 4.2 years (HR 1.08, 95% CI 0.73 to 1.60, p=0.69), but the risk of
any stroke was greater in the CAS group (HR 1.73, 95% CI 1.29 to 2.32, p<0.01)
reflecting an excess of minor perioperative stroke in patients undergoing CAS [83].
39
Finally, the North American Carotid Revascularization: Endarterectomy versus Stent
Trial (CREST) began recruiting symptomatic patients, but later amended its protocol to
allow randomization of asymptomatic patients. CREST enrolled 2502 patients in total.
Symptomatic patients were required to have ≥50% stenosis on angiography, ≥70%
stenosis on Doppler ultrasound or ≥70% stenosis on MRA or CTA if ultrasound
indicated 50-69% stenosis. Asymptomatic patients were required to have ≥60%
stenosis on angiography, ≥70% on ultrasound, or ≥80% on MRA or CTA if ultrasound
indicated 50-69% stenosis. CREST also differed from the European trials in mandating
the use of a single manufacturer’s devices for CAS – the Acculink stent and the
Accunet CPD. No significant difference was found between CAS and CEA for the
primary outcome measure of any periprocedural stroke, MI or death or postprocedural
ipsilateral stroke at four years follow-up (7.2% vs 6.8%, HR 1.11, 95% CI 0.81 to 1.51),
but there was a higher risk of periprocedural stroke in the CAS group (4.1% vs 2.3%,
p=0.01) in keeping with the results of EVA-3S, SPACE and ICSS [84]. This analysis
included both asymptomatic and symptomatic patients.
The short-term results of symptomatic patients only from these trials are summarized in
Table 2.
40
Table 2. Results of the recent large randomized controlled trials of CAS vs CEA in symptomatic patients
Trial Region Sample
Size (n)
Comparison Outcome Risk of outcome at 30 days post-procedure
EVA-3S
[79]
France 527 CAS + CPD vs
CEA
Any periprocedural stroke or
death
9.6% vs 3.9% (p=0.01) by intention to treat
SPACE
[77]
Germany,
Austria,
Switzerland
1183 CAS +/- CPD vs
CEA
Any periprocedural ipsilateral
ischemic stroke or death
6.8% vs 6.3% (p=0.09 for non-inferiority) by
intention to treat
ICSS
[82]
International 1713 CAS +/- CPD
vs
CEA
Any periprocedural stroke, MI or
death
7.4% vs 4.0% (p<0.01) per protocol
CREST
(symptomatic
patients only)
[85]
USA,
Canada
1321 CAS + CPD vs
CEA
Any periprocedural stroke, MI or
death
6.7% vs 5.4% (p=0.30) by intention to treat
41
12.4.2 Trials of CAS vs CEA – asymptomatic stenosis
The safety and efficacy of CAS against CEA has also been tested in asymptomatic
patients. Three large trials have reported results. The largest cohort of asymptomatic
patients is found in the CREST study, which enrolled 1181 patients with carotid
stenosis >50% but without symptoms. The combined risk of stroke, myocardial
infarction (MI) or periprocedural death within 30 days of the procedure in these
asymptomatic patients was 2.5% in the CAS group and 1.4% in the CEA group (HR
1.88, 95% CI 0.79 to 4.42, p=0.15) [85] suggesting that stenting was more risky than
endarterectomy although the difference was not statistically significant. The results of
long-term follow-up are awaited. In the older Stenting and Angioplasty with Protection
in Patients at High Risk for Endarterectomy (SAPPHIRE) study, 7 out of 117
asymptomatic patients in the CAS group and 5 out of 120 asymptomatic patients in the
CEA group suffered stroke or periprocedural death [86]. Brooks et al. also enrolled 85
asymptomatic patients with carotid stenosis >80% in their trial of CAS vs CEA, but
none suffered stroke or death following the procedure [71].
Another study conducted in China, the Trial of Endarterectomy versus Stenting for the
treatment of Carotid Atherosclerotic Stenosis (TESCAS-C) [87], enrolled a mixture of
symptomatic and asymptomatic patients, but to date has not reported these results
separately. And one further trial, the North American Asymptomatic Carotid Trial (ACT-
1), was designed to compare CAS with CEA by randomizing patients in a 3:1 ratio to
these treatments. Its protocol specified exclusive use of the Xact carotid stent and
Emboshield protection device in the CAS arm, but has now stopped recruitment as a
result of a decision by the sponsor to discontinue funding [88].
Meta-analysis of these trials suggests a trend towards worse outcome with CAS in
asymptomatic patients in the short-term with a higher risk of periprocedural death or
stroke, and in the long-term with a higher risk of periprocedural death or stroke or
ipsilateral stroke thereafter. However, the total number of patients enrolled in
asymptomatic trials is low, and this excess risk does not reach statistical significance
[86].
12.4.3 Cost considerations
Limited research exists on the potential health economic benefits of carotid stenting
versus endarterectomy. In theory, stenting may offer a reduced hospital stay,
avoidance of a general anaesthetic and, at least in younger patients, a similar
42
complication rate to that of CEA while avoiding cranial nerve palsy or haematoma from
a neck incision.
The data on the cost of the procedure itself is mixed – some studies find CAS to be
more expensive [89], while other analyses attribute a higher cost to CEA [90]. Of
interest, data from the USA showed that the net revenue for hospitals was 29% higher
when patients were treated with CEA rather than CAS [91]. In order to be as cost-
effective as surgery, CAS operators and device manufacturers will need to consistently
demonstrate a similar complication rate to that of CEA in low-risk patients, reduce the
cost of the procedure, or perhaps restrict the procedure to patients at higher surgical
risk [92]. A cost-effectiveness analysis in the International Carotid Stenting Study is
ongoing.
From trial data to individual patients: predicting procedural risk 12.5
Complications following any surgical procedure remain common, and occur in up to 7%
of all patients even when specific interventions are applied to reduce their risk [93].
Examples of postoperative problems include cardiac arrest, myocardial infarction,
pulmonary embolism, bleeding, pneumonia, wound problems including infection, sepsis
and death. Data suggest that up to half of these events are potentially avoidable [94]
[95]. Complication rates in carotid artery surgery and stenting remain significant [96]
[97], and excluding those patients who would suffer stroke, myocardial infarction or
other serious complications would prevent much perioperative morbidity and mortality.
Studies identifying subgroups of patients at risk of complications following CEA and
CAS are reviewed in detail in Chapters 14, 15 and 17. However, an overview of
important risk factors that appear to have a strong influence on outcome and that have
been reported in multiple studies or meta-analysis is presented here.
12.5.1 Predicting the risk of CEA
Patient factors
Risk modelling drawing on patient data from ECST suggests that there are specific
subgroups of patients at increased risk of stroke or death following CEA. These include
female patients, those with peripheral vascular disease, and those with a high systolic
blood pressure before surgery [98].
43
Some anatomical factors – previous surgery, radiotherapy, restenosis of a vessel, a
high stenosis or inflexible neck – are thought to confer high risk for carotid
endarterectomy because of limited surgical access to the carotid plaque or a
technically difficult procedure, but there is some evidence from small case series to
suggest that patients with some of these factors may not actually suffer more
complications [99] [100]. In addition, the patient with heart or lung disease is often
thought to be at increased risk from surgery [101], but again this is not always borne
out in case series [100] [102].
Technical variables
CEA can be performed under local or general anaesthesia with seemingly little
difference in outcome [103] [104], but with the advantage of clinical monitoring of the
patient for signs of focal neurological dysfunction during the procedure if local
anaesthetic is used. Contralateral ICA stenosis appears to increase the perioperative
risk of stroke [105], and the use of shunts during surgery to bypass arterial clamping on
the ipsilateral side and maintain distal flow to the brain has become widespread despite
limited evidence of efficacy [106]. After resection of the atherosclerotic plaque the
vessel may be repaired by primary closure, by patch or by vein graft. The safety of
differing surgical techniques appears similar, although there is some evidence that
patch closure may reduce late restenosis [107]. Figure 5 illustrates the process of
exposing the atheromatous plaque in eversion carotid endarterectomy – one surgical
technique for performing arteriotomy and removal of the plaque.
44
Figure 5. Surgical technique for eversion carotid endarterectomy
a. Haemostatic control is achieved by clamps on common, internal and external
carotid arteries
b. Arteriotomy is made, extending toward the fork of the carotid bifurcation
c. Complete transection of the internal carotid artery is followed by resection of the
plaque then surgical repair of the artery
From [108] with kind permission from Springer Science and Business Media
45
Centre and operator experience
Technical errors made during surgery may contribute to perioperative stroke [109].
Several studies address the influence of operator experience or volume in carotid
endarterectomy. Surgeons performing a high volume of operations appear to have
lower mortality and postoperative stroke rates [110] [111] [112]. There is also a
relationship between the annual hospital volume of endarterectomies and adverse
neurological outcome [113], and UK professional bodies are moving to support
centralization of vascular surgical services in high-volume centres [114].
12.5.2 Predicting the risk of CAS
Patient factors
There appears to be a strong and consistent effect of the patient’s age on the outcome
of stenting, with a doubling of the risk of any stroke or death up to 120 days after CAS
compared with CEA in patients over the age of 70 years (RR 2.04, 95% CI 1.48 to
2.82) in meta-analysis of the results of the European trials of CAS versus CEA [115].
CREST separately reported a strikingly similar finding – the efficacy of CAS and CEA
was approximately equal at age 70, but the risk of CAS rose with increasing age while
there was no corresponding increase in risk in the CEA group [84] [116]. There is also
recent evidence that the patient with a higher burden of white matter brain disease at
baseline has an increased risk of stroke following CAS [117], independent of the
patient’s age.
CREST reported a worse outcome after CAS in women – the 30-day risk of stroke, MI
or death was 6.8% in women randomized to CAS versus 3.8% in women randomized
to CEA (HR 1.84, 95% CI 1.01 to 3.37), whereas there was no significant difference
between CAS and CEA complications in men. However, there was no difference in
long-term outcome in CREST between the sexes and the Carotid Stenting Trialists’
Collaboration did not report a difference in short-term outcomes between men and
women in the European trials of CAS versus CEA [118] [84].
Technical variables
A variety of endovascular devices from different manufacturers are available to
interventionists. While some clinical trials such as the International Carotid Stenting
Study approved a wide range of devices [81], others including CREST mandated
specific devices [84]. Thus evidence comparing devices to each other is sparse.
46
However, by grouping devices by their characteristics some distinction can be made
between the performance of closed-cell and open-cell stents. Closed-cell stents offer
enhanced coverage of the atheromatous lesion with a smaller area between metallic
struts, and there is a suggestion that they offer better protection against stroke and TIA
by containing debris at the site of the lesion [119].
Anatomical features
Several vascular anatomical features are thought to confer increased risk of stroke
following CAS. Among those that have been shown to be associated with an increased
risk are ICA angulation ≥60o [120] [121], lesion length >15mm, type III or “bovine” aortic
arch [122], aortic calcification [122], a calcified or ulcerated lesion [122] or an ostial
lesion [122] [121]. In addition to these features, an angulated distal ICA and “pinhole”
tight stenosis are described as technically challenging [123]. If some vascular
anatomical variants are indeed more difficult to navigate and stent than others, it
follows that more experienced operators could be able to avoid dislodging thrombus or
fragmenting atheroma and therefore avoid perioperative stroke.
Centre and operator experience
While there appears to be remarkable consistency across the results of these
randomized trials, the design of the European trials (ICSS, EVA-3S and SPACE) in
particular have been criticized. As CAS is the newer procedure, the number of
interventionists able to demonstrate substantial volumes of stenting procedures is
smaller, and this was reflected in the criteria for centre enrolment in ICSS, for example.
Surgeons performing endarterectomy in ICSS were required to have performed a
minimum of 50 prior CEA procedures with a minimum volume of 10 cases per year and
demonstrate acceptable patient outcome data. In contrast, interventionists could
perform CAS procedures in the trial having placed only 10 carotid stents (albeit with a
minimum of 50 stenting procedures in total including those in other vascular territories).
Assessing the risk of complications with any individual operator is difficult, and larger
numbers of patients are required to report a risk with narrow confidence intervals [124].
This has led to the suggestion that inexperience in these trials is responsible for the
higher rate of stroke in CAS patients, and that perhaps “CAS performed by
inexperienced interventionists has higher peri-procedural complication rates than CEA
done by experienced surgeons” [125] is the most valid conclusion to be drawn from the
data from these trials. Recent pooled analysis of the 1546 patients undergoing CAS in
the three European trials suggests that the patients of interventionists who performed a
47
low number of procedures each year within the trials were at double the risk of stroke
or death within 30 days of CAS compared with the patients of interventionists who
performed more than 5 stenting procedures within the trial each year (RR 2.30, 95% CI
1.36 to 3.87) [126].
The current status of CAS & CEA for symptomatic patients 12.6
Carotid revascularization is recommended in UK, US and European guidelines for
patients with symptomatic stenosis ≥50% where the operative morbidity is estimated to
be low. 2014 AHA guidelines note the reduced benefit of CEA in patients with 50-69%
stenosis, and suggest that the intervention in these patients should be “performed by a
surgeon with excellent operative skills” [127]. In these guidelines, CAS is deemed to be
indicated as an alternative to CEA in patients at average or low risk of complications,
taking age into account, or in those with contraindications to CEA. In the UK the
National Institute for Clinical Excellence (NICE) guidelines from 2008 do not specify a
recommendation for the use of CAS other than in cases of symptomatic carotid
stenosis ≥50% where endarterectomy might be high risk or contraindicated [128] [129].
Similarly, the European Vascular Society guidelines continue to recommend CEA over
CAS for the majority of symptomatic patients [46] [130].
Are these guidelines reflected in clinical practice? Data from hospitals in England
reveals that the overwhelming majority of revascularization procedures carried out on
the carotid artery are endarterectomies, with only 5% of patients receiving stenting
[131]. The results of the major randomized trials that overall favour CEA as the safer
procedure – SPACE, EVA-3S, ICSS and CREST – appear not to have affected the
number of CAS procedures being carried out between 2006 and 2012, but interestingly
the patients undergoing CAS were younger in age, perhaps reflecting the
recommendation that CAS be considered more carefully in older patients because of
the higher risk of peri-procedural complications. In the United States, there has been a
significant increase in the number of CAS procedures performed since the publication
of the main results of these trials predominantly, again, in younger patients and those
who were asymptomatic [132]. The debate as to which procedure is superior seems far
from settled in clinicians’ minds.
This thesis 12.7
This thesis will examine the risk of, and impact of, complications arising from carotid
revascularization in symptomatic patients in the International Carotid Stenting Study –
48
stroke, myocardial infarction, death, cranial nerve palsy and neck haematoma. It will
determine whether there are subgroups of patients at particular risk of these
complications following CEA or CAS based on their demographic characteristics, the
type of procedure they have undergone or their individual vascular anatomy. Finally it
will examine the performance of existing risk-scoring systems for predicting
complications of CEA and CAS when applied to the patients in ICSS who were
randomized to those procedures. The thesis concludes with a summary and discussion
of its main findings, and places the work in the context of recent and ongoing studies of
CEA and CAS.
49
13. Methods – ICSS and the ICSS-MRI Substudy
The International Carotid Stenting Study 13.1
13.1.1 Introduction
The full protocol for ICSS is presented in Appendix IV, and is published elsewhere [81].
The ICSS-MRI Substudy protocol is presented in full in Appendix V. The International
Carotid Stenting Study was conceived after the conclusion of CAVATAS [75], which
found no statistically significant difference in outcome between CAS and CEA
immediately following treatment or in the long term. However, carotid angioplasty
continued to evolve during the trial, and the use of stents in the procedure developed
while CAVATAS was still recruiting. CAS still promised the advantages of reduced
hospital stay, potential avoidance of a general anaesthetic, and a reduced rate of
surgery-specific complications such as neck haematoma and cranial nerve palsy. ICSS
was therefore designed as an international, multicentre, randomized, controlled open
clinical trial to test the difference between the long-term risk of fatal or disabling stroke
in patients with symptomatic carotid artery stenosis greater than 50% after
randomization to either CEA or CAS.
13.1.2 Selection and enrolment of trial centres & interventionists
50 academic centres in Europe, North America and Oceania enrolled patients in ICSS.
Each was required to provide a principal investigator qualified in neurology or internal
medicine to see patients prior to randomization and then again in follow-up. Surgeons
performing trial procedures were required to have carried out 50 CEA operations with
an annual rate of at least 10 cases per year, while CAS interventionists were required
to have performed 50 stenting procedures, at least 10 in the carotid artery territory. All
were expected to provide audit data to an ICSS accreditation committee showing a
satisfactory complication rate similar to that of ECST (7.0% risk stroke or death within
30 days of the procedure, 95% CI 5.8% to 8.3%) [36].
Centres unable to fulfil the above requirements for CAS competency were allowed to
join the trial and enrol patients as a “probationary centre”. CAS procedures carried out
at these centres were supervised by an experienced interventionist until 10 more
successful procedures were completed within the trial and the proctor was satisfied that
the interventionist in question could carry out the procedure independently.
50
13.1.3 Selection of patients and randomization
Patients randomized in ICSS had symptomatic extracranial stenosis of the carotid
artery attributable to atherosclerotic disease. Stenosis had to be greater than 50% as
measured by the NASCET method [40] or an equivalent non-invasive method such as
carotid ultrasound, and symptoms were required to have occurred within the last 12
months.
Patients were excluded if they were thought unsuitable for either treatment due to
Tortuous carotid vascular anatomy
Proximal common carotid artery disease or the presence of thrombus
Pseudo-occlusion (>99% stenosis with normal distal runoff on angiography)
Distal stenosis or rigid neck rendering surgery hazardous or impossible
Previous CAS or CEA in the artery to be randomized
Clinical instability or serious co-morbid illness (“not fit for surgery”)
A life expectancy of less than 2 years due to a pre-existing condition
Planned coronary artery bypass grafting within 1 month of CAS or CEA
Refusal or inability to give written informed consent to trial participation or
refusal to be randomized to either treatment
Major stroke with no useful recovery of function within the territory of the artery
to be randomized
Prior to randomization patients were given written information about the trial and
informed consent to participate was given by the patient.
13.1.4 Baseline visit and investigations
At baseline visit, investigators recorded demographic information about the patient and
details of past medical history including vascular risk factors. Brain imaging, in the form
of CT or MRI, was required to identify previous areas of infarction or haemorrhage and
to provide a comparison with any post-procedure scans.
To confirm the presence and extent of carotid stenosis, imaging from at least one of the
following modalities was required
Catheter angiogram showing the carotid bifurcation
Bilateral MRI angiograms with a concordant Doppler ultrasound scan
51
Bilateral CT angiograms with a concordant Doppler ultrasound scan
Bilateral Doppler ultrasound scan only, if the centre was able to provide
documented evidence of the reliability of their results through clinical audit and
if it was standard practise to treat on the basis of ultrasound alone at that centre
Pre-randomization clinical and imaging data were returned to the trial office in London.
13.1.5 Randomization procedure
Randomization was carried out by the Oxford Clinical Trials Service Unit, stratified by
centre with risk factors balanced between the two arms. Minimization was used to
balance patient characteristics between the two arms. Patients with stenosis in both
carotid arteries could only be randomized once, for the symptomatic artery to be
treated first. The allocated procedure was then required to be carried out as soon as
safely and practically possible after randomization. The possibility of a patient
becoming ineligible for revascularization after randomization in the trial or crossing over
to the other treatment was anticipated, and these patients were included in intention-to-
treat analyses.
13.1.6 Carotid endarterectomy
Carotid endarterectomy was carried out according to the operating surgeon and the
trial centre’s usual protocol. The type of anaesthesia to be used was not specified nor
was the choice of arterial reconstruction. The use of a patch to close the arteriotomy
was optional. All patients undergoing surgery were required to receive “best medical
care” including management of vascular risk factors and appropriate antiplatelet,
antihypertensive or cholesterol-lowering drugs.
13.1.7 Carotid stenting
Devices used in the carotid stenting procedure were required to be CE-marked and
were approved by the trial steering committee. The use of a cerebral protection device
was optional, but recommended “whenever the operator [thought] one could be safely
deployed”. Aspirin and clopidogrel prescription prior to the procedure and for a
minimum of 4 weeks after was recommended. The use of intra-procedural heparin to
prevent acute thrombosis secondary to instrumentation of the carotid was mandatory,
as was the administration of atropine or equivalent drugs to prevent haemodynamic
disturbance secondary to manipulation of the carotid baroreceptors. Finally, it was
anticipated that most patients would require balloon dilation (“pre-dilation”) of the
52
stenotic lesion to facilitate the passage of equipment through the stenosis. Patients
undergoing CAS otherwise received the same “best medical care” as the surgical
group. Technical details of either CAS or CEA were returned to the trial office.
13.1.8 Post-procedure follow-up
Patients were monitored in hospital, and events occurring between the procedure and
discharge were reported on the technical data form. Patients were then followed-up by
a trial investigator at one month after the procedure.
13.1.9 Outcome events and adjudication
The main outcome events examined in the primary outcome of the trial and
subsequently in this thesis are defined below
Stroke – “an acute disturbance of focal neurological function with symptoms
lasting more than 24 hours resulting from intracranial vascular disturbance”
This diagnosis included retinal stroke causing loss of vision in one eye for more
than 24 hours
Myocardial infarction – recorded if the patient experienced two out of
o Cardiac enzyme rise more than twice the upper limit of normal range
o The development of specific ECG abnormalities
o A history of chest discomfort for at least 30 minutes
Cranial nerve palsy (within 30 days of the procedure) – reported by the
investigator in individual patients who experienced motor or sensory
disturbance, attributed to the procedure, in one of the cranial nerves.
Haematoma – defined by the investigator in patients who experienced bleeding
in the neck as a result of revascularization. “Severe haematoma” was defined
as haematoma requiring blood transfusion, re-operation or one that prolonged
hospital stay
Death
A clinical diagnosis of stroke was confirmed by brain CT or MRI carried out as soon as
possible after the clinical event. The cause of death was confirmed through clinical
reports from the treating centre and death certificate or post-mortem results where
available.
53
Potential stroke or MI outcome events were reported in detail to the trial office as
above. An adjudication process was carried out to decide if the reported event met the
criteria of one of the trial outcome events. Adjudication was carried out by two
investigators at the trial office blinded to allocated treatment. A third, independent,
adjudicator then reviewed the event. If a disagreement occurred between adjudicators
the difference was resolved with the opinion of a second independent adjudicator and
consensus. Additional information on the clinical consequences of cranial nerve palsy
occurring within 30 days of CEA was sought by means of a questionnaire, detailed in
Appendix II.
Stroke or cranial nerve palsy was flagged as “disabling” if the patient experienced an
increase in their modified Rankin Scale (mRS) score to 3 or more where that increase
was attributable to the event at a specified time-point during follow-up.
13.1.10 Sample size and statistical analysis
Sample size in ICSS was planned as 1500 patients in fully-enrolled centres, with a
calculated 95% confidence interval for a difference in risk between CAS and CEA of
±3% for the composite outcome of stroke, myocardial infarction or death within 30 days
of treatment. Additional patients were enrolled in probationary centres.
13.1.11 Ethical approval and safety monitoring
ICSS was approved by a Multicentre Research Ethics Committee (MREC) in the United
Kingdom. In addition, individual centres obtained their own local ethics committee
approval before starting randomization in the trial.
The safety of the patients enrolled in ICSS was overseen by a Data Monitoring
Committee comprising a neurologist, medical statistician, surgeon and interventionist.
Interim analyses were carried out during the trial to ensure that morbidity and mortality
remained within acceptable limits. Individual centres’ results were monitored by the trial
manager, with two consecutive deaths or three consecutive major events at an
individual centre within 30 days of treatment in one arm of the study prompting a review
of outcome events at this centre. Similarly, a cumulative stroke, MI or death risk of
>10% over 20 cases would prompt a review.
During the course of the trial concern was raised, by the above mechanism, about the
results of two investigators performing CAS procedures at two separate supervised
centres. Both centres suspended recruitment while outcome events were reviewed,
54
although one centre subsequently re-started enrolment into ICSS with a different
interventionist performing trial procedures [82]. The results presented in this thesis
include patients treated by these investigators where the patients received their
allocated treatment.
The ICSS-MRI Substudy 13.2
13.2.1 Introduction
Complementary to the International Carotid Stenting Study was the ICSS-MRI
Substudy. The aim of this substudy was to compare the risk of ischaemic (and possibly
sub-clinical) brain injury as assessed by MRI in patients undergoing stenting in ICSS
compared to those undergoing surgery.
13.2.2 Substudy design
Patients enrolled in the ICSS-MRI Substudy received standard treatment as specified
in the ICSS protocol above. They additionally underwent brain MRI 1-3 days before
treatment, 1-3 days after treatment and approximately 30 days after revascularization.
Patients with a contraindication to MRI (e.g. metallic implant or claustrophobia) were
excluded from participation in the study. The following MRI sequences were specified
in the trial protocol
Diffusion-weighted imaging (DWI) to detect acute brain ischaemia or infarction
Gradient echo T2*-weighted to detect brain haemorrhage
T1-weighted and fluid-attenuated inversion recovery (FLAIR) to assess whether
acute brain lesions on DWI persisted to become permanent FLAIR lesions at 30
days after the revascularization
13.2.3 Sample size, data and statistical analyses
A sample size of 200 patients was planned assuming a risk of new DWI lesions after
endarterectomy of 25%, giving a 90% power to detect a doubling of the risk of a new
DWI lesion in the stenting group.
The analysis of brain imaging in the ICSS-MRI Substudy is described in more detail in
Chapter 18. Two investigators independently recorded the presence, volume and
location of ischaemic and haemorrhagic lesions on each scan, and a third investigator
55
reviewed in cases of disagreement. Scans were reported blind to allocated treatment or
the nature of the scan (e.g. pre- or post-procedural).
CONSORT flow diagram 13.3
The progress of patients through the International Carotid Stenting Study is outlined in
the CONSORT flow diagram in Figure 6 below.
Trial funding 13.4
ICSS was funded by the UK Medical Research Council, the Stroke Association, Sanofi-
Synthélabo, and the European Union. ICSS is a registered clinical trial: ISRCTN
25337470.
56
Figure 6. ICSS CONSORT trial diagram
855 assigned to Stenting 858 assigned to Endarterectomy
853 analysed by intention to treat from
randomisation until end of follow-up
857 analysed by intention to treat from
randomisation until end of follow-up
2 withdrew all consent
immediately after randomisation
1 withdrew all consent immediately
after randomisation
9 crossed over to endarterectomy
2 anatomy unsuitable
3 medical contraindications
1 refused treatment
3 other reasons
16 underwent no procedure
1 disabling stroke before
intended procedure
5 artery occluded
3 artery less than 50% stenosed
1 anatomy unsuitable
2 other medical contraindications
3 other reasons
15 crossed over to stenting
1 anatomy unsuitable
6 medical contraindications
4 refused treatment
4 other reasons
21 underwent no procedure
2 died before intended procedure
3 disabling stroke before intended
procedure
9 artery occluded
1 artery less than 50% stenosed
3 medical contraindications
1 refused treatment
2 other reasons
828 stenting procedures initiated 821 endarterectomy procedures initiated
64 stenting aborted
752 per-protocol patients
reached 30-day follow-up in
ICSS
12 follow-up <30 days
10 fatal events
2 had no further follow-up
2 endarterectomy aborted
811 per-protocol patients
reached 30-day follow-up in
ICSS
8 follow-up <30 days
4 fatal events
4 had no further follow-up
1713 patients randomised
57
14. Risk factors for stroke, myocardial infarction or death
following carotid endarterectomy
Introduction 14.1
The North American Symptomatic Carotid Endarterectomy Trial (NASCET) [35], the
European Carotid Surgery Trial (ECST) [36] and the smaller Veteran’s Affairs Trial [39]
demonstrated a reduction in the long-term rate of recurrent stroke in patients with
>50% symptomatic carotid artery stenosis undergoing carotid endarterectomy
compared with those on medical therapy. Meta-analysis of the results of these three
trials, and two smaller studies, suggests that for every 11 symptomatic patients
undergoing CEA one outcome event is prevented for the combined endpoint of any
stroke or perioperative death [133].
However, CEA in these trials carried a risk of perioperative stroke or death of 7.0%
(95% CI 6.2 to 8.0) [133] and subsequent studies, focussing on these short-term
measures of safety, continue to demonstrate a significant risk of major complications
following revascularization [134] [96]. Perioperative stroke and MI have been shown to
have a significant adverse impact on long-term survival, and stroke in particular may
halve survival in the first year after surgery [135]. The choice of anaesthesia and
surgical technique during CEA varies [136] [137], and despite CEA being a long-
established technique for revascularization there remains debate about the optimal
processes of care for surgery including perioperative medical therapy, type of arterial
reconstruction and mode of anaesthesia to prevent complications.
This analysis aimed to determine whether there were groups of patients at higher risk
of stroke, MI or death within 30 days of CEA in the International Carotid Stenting Study
and whether there are surgical practices associated with higher risk.
Methods 14.2
14.2.1 Study design & CEA procedure
This chapter presents a per-protocol analysis of patients enrolled in ICSS who were
randomized to CEA and in whom the allocated procedure was initiated. The inclusion
and exclusion criteria for ICSS are outlined in Chapter 13.1 and the trial protocol is
detailed in full in Appendix IV. Patients over the age of 40 years old could be enrolled in
ICSS if they presented with recently-symptomatic carotid artery stenosis >50% due to
58
atheroma. Patients were excluded from randomization in ICSS if they were unsuitable
for surgery due to a surgically-inaccessibly high level of stenosis, if they had a rigid
neck making positioning difficult for surgery, or if there was previous revascularization
of the artery to be randomized.
CEA in ICSS was performed according to the surgeon’s usual practice – the type of
arterial reconstruction was not specified, nor was the mode of anaesthesia. Medical
therapy for control of vascular risk factors was required in both groups in the trial, but
the prescription of one or more antiplatelet agents prior to or after surgery was at the
operator’s discretion. Surgeons in ICSS were required to provide audited complication
rates within acceptable limits and to have performed a minimum of 50 CEA procedures
prior to joining the trial with a minimum volume of 10 procedures per year. In addition to
this technical information, centres provided demographic information about patients
and specified whether their general policy was to send patients to a specialized post-
procedure ward such as a high-dependency or intensive care unit, or whether patients
were admitted to a general surgical or medical ward post-operatively.
14.2.2 Outcome events
The combined endpoint of stroke, MI or death within 30 days of the procedure was
analysed. Events were reported in detail to investigators at the central trial office.
Confirmatory evidence was required where available: CT or MRI brain after stroke,
cardiac enzymes and / or ECG after MI, death certificate and autopsy report if available
following death. Stroke was defined as “an acute disturbance of focal neurological
function lasting more than 24 hours resulting from intracranial vascular disturbance”. A
diagnosis of MI required two out of three features – a clinical history of chest discomfort
lasting at least 30 minutes, cardiac enzyme rise more than twice the upper limit of
normal or the development of specific ECG abnormalities. Outcome events were
submitted to an external, independent, adjudicator who was masked to treatment
allocation. If their assessment differed from the trial office’s initial assessment a second
external adjudicator reviewed the event. Differences in adjudication were resolved by
consensus.
14.2.3 Statistical methods
Only patients in ICSS in whom the allocated procedure was initiated were included in
this analysis. Trial patients who crossed-over to medical therapy or received CEA after
a prior stenting procedure (attempted or completed) were excluded from this analysis.
59
CEA was considered “initiated” if the patient underwent either local or general
anaesthesia prior to surgical incision. Risk factors for stroke, MI or death were
examined sequentially in a univariable binomial regression analysis using maximum
likelihood estimation. Second events within 30 days of the procedure were not counted.
Patients with missing data were excluded from each relevant analysis. The risk ratio for
each baseline or technical variable was estimated with a 95% confidence interval. Wald
tests were used for continuous and binary predictors, with an overall likelihood test for
categorical predictors of more than two levels. No correction was applied for multiple
comparisons. p<0.05 was accepted as conferring statistical significance in all analyses.
A multivariable model was developed using a forward stepwise approach. Analyses
were performed with Stata (StataCorp, 2011. Stata Statistical Software: Release 12.
College Station, TX: StataCorp LP).
Results 14.3
14.3.1 Outcome events
14.3.2 Patient and procedure characteristics
CEA was initiated in 821/855 patients allocated to the procedure in ICSS. Patient and
procedural characteristics for this group are presented in detail in Figure 7 along with
the results of univariable analysis. 70.4% of patients were male and 52% of patients
were aged 70 years or more (the median age in the CEA arm of ICSS at the date of
enrolment). Vascular risk factors were common. 21.2% of patients were diabetic,
66.2% were receiving treatment for hyperlipidaemia and 69.7% were receiving
treatment for hypertension.
Arterial reconstruction was by patch closure in 55.9%, “standard” closure without patch
in 22.1% and by eversion endarterectomy in 6.0%. Vein interposition was used in only
0.4%. General anaesthesia or combined local-general anaesthesia was administered
to 79.2% of patients. A shunt was used during the procedure in 39.5%. An antiplatelet
drug was prescribed to 88.4% of patients prior to the procedure, of whom 30.1% were
taking two or more antiplatelet medications.
14.3.3 Characteristics and timing of outcome events
27/821 patients (3.3%) suffered a stroke of any severity within 30 days of the
procedure, 5 suffered MI, and one patient died of another cause giving a combined risk
of stroke, death or MI within 30 days of the procedure of 4.0%. 13/33 (39.4%) of these
60
events occurred on day 0 – the day of the procedure, the remainder occurring between
days 1 and 30. 7/33 (21.2%) events occurred on or after the date of discharge. The
median length of stay was 4 days before discharge.
14.3.4 Patient and procedural risk factors for stroke, MI or death within 30 days
of the procedure
The results of univariable analyses of patient and procedural factors on the risk of
stroke, MI or death within 30 days of CEA are presented in Figure 7. The risk of the
combined outcome was significantly higher in female patients (RR 1.98, 95% CI 1.02 to
3.87, p=0.05) and with each increasing 10mmHg of baseline diastolic blood pressure
(RR 1.31 for each +10mmHg, 95% CI 1.20 to 1.66, p=0.04). The time from the index
event that prompted randomization in the trial (stroke or TIA) to the date of surgery was
also a significant predictor, but this this result was influenced by one outlying patient.
Removing this patient from the per-protocol population in a sensitivity analysis resulted
in a non-significant p-value (RR 1.01 per 7 days, 95% CI 0.97 to 1.06, p=0.48). The
age of the patient, the side of surgery, the level of disability and the degree of ipsilateral
or contralateral carotid artery stenosis were not predictors of the risk of stroke, MI or
death within 30 days of surgery.
Shunt use (RR 0.99, 95% CI 0.50 to 1.96, p=0.98), local anaesthesia only vs general or
combined local-general anaesthesia (RR 0.48, 95% CI 0.15 to 1.57, p=0.23) and the
type of surgical reconstruction (i.e. primary closure vs patch vs vein interposition) did
not statistically significantly influence risk despite one out of three patients in whom
vein interposition was used experiencing an event.
Centres with a policy of sending patients to a specialized ward (such as a high
dependency or intensive care unit) following CEA reported a lower risk of stroke, MI or
death within 30 days of surgery (RR 0.57, 95% CI 0.29 to 1.13, p=0.11), but this result
was not statistically significant.
61
Figure 7. Univariable predictors of the risk of stroke, myocardial infarction or death
in 821 patients undergoing CEA in whom the procedure was initiated
Reconstruction:StandardEversionPatchVein Interposition
Shunt use:NoYes
Anaesthesia:GeneralLocal
Antiplatelet pre-procedure:NoYes
>1 antiplatelet pre-procedure:NoYes
Anticoagulant pre-procedure:NoYes
Antiplatelet post-procedure:NoYes
Anticoagulant post-procedure:NoYes
Sex:MaleFemale
Age in years:<7070+
Treated hypertension:NoYes
Cardiac failure:NoYes
Angina:NoYes
Previous MI:NoYes
Previouis CABG:NoYes
Atrial fibrillation:NoYes
Other cardioembolic source of embolus:NoYes
Diabetes:NoYes
Variable
10/182 (5.5%)1/49 (2%)13/459 (2.8%)1/3 (33.3%)
20/494 (4%)13/324 (4%)
28/650 (4.3%)3/144 (2.1%)
2/52 (3.8%)27/726 (3.7%)
17/531 (3.2%)12/247 (4.9%)
22/619 (3.6%)7/159 (4.4%)
5/87 (5.7%)24/691 (3.5%)
23/574 (4%)6/204 (2.9%)
18/578 (3.1%)15/243 (6.2%)
13/391 (3.3%)20/430 (4.7%)
6/242 (2.5%)27/572 (4.7%)
33/770 (4.3%)0/44 (0%)
27/742 (3.6%)6/72 (8.3%)
25/664 (3.8%)8/150 (5.3%)
31/703 (4.4%)2/111 (1.8%)
32/760 (4.2%)1/54 (1.9%)
31/798 (3.9%)2/16 (12.5%)
24/640 (3.8%)9/174 (5.2%)
/ No. pts (%)No. events
1.000.37 (0.05,2.83)0.52 (0.23,1.15)6.07 (1.10,33.54)
1.000.99 (0.50,1.96)
1.000.48 (0.15,1.57)
1.000.97 (0.24,3.96)
1.001.52 (0.74,3.13)
1.001.24 (0.54,2.85)
1.000.60 (0.24,1.54)
1.000.73 (0.30,1.78)
1.001.98 (1.02,3.87)
1.001.40 (0.71,2.77)
1.001.90 (0.80,4.55)
N/A
1.002.29 (0.98,5.36)
1.001.42 (0.65,3.08)
1.000.41 (0.10,1.68)
1.000.44 (0.06,3.16)
1.003.22 (0.84,12.31)
1.001.38 (0.65,2.91)
ratio (95% CI)Risk
0.12
0.98
0.23
0.96
0.26
0.61
0.29
0.49
0.05
0.34
0.15
0.06
0.38
0.22
0.41
0.09
0.40
P-value
1.000.37 (0.05,2.83)0.52 (0.23,1.15)6.07 (1.10,33.54)
1.000.99 (0.50,1.96)
1.000.48 (0.15,1.57)
1.000.97 (0.24,3.96)
1.001.52 (0.74,3.13)
1.001.24 (0.54,2.85)
1.000.60 (0.24,1.54)
1.000.73 (0.30,1.78)
1.001.98 (1.02,3.87)
1.001.40 (0.71,2.77)
1.001.90 (0.80,4.55)
N/A
1.002.29 (0.98,5.36)
1.001.42 (0.65,3.08)
1.000.41 (0.10,1.68)
1.000.44 (0.06,3.16)
1.003.22 (0.84,12.31)
1.001.38 (0.65,2.91)
ratio (95% CI)Risk
0.12
0.98
0.23
0.96
0.26
0.61
0.29
0.49
0.05
0.34
0.15
0.06
0.38
0.22
0.41
0.09
0.40
P-value
1.05 .2 .5 1 2 4 6 10 15
Risk Ratio (95%CI)
62
Figure 7 continued
Peripheral arterial disease:NoYes
Smoking status:NeverFormerCurrent
Treated hyperlipidaemia:NoYes
Degree of stenosis in treated artery:50-69%70-99%
Degree of stenosis in contralateral artery:0-49%50-69%70-99%Occluded
Nature of ipsilateral index event:StrokeRetinal StrokeTIAAFX
Multiple ipsilateral events prior to randomisation:NoYes
Prior ipsilateral stroke:NoYes
Baseline mRS:012345
Side of procedure:LeftRight
Centre policy post-procedure:WardICU
Clamping time (per 20 mins):
Duration of surgery (per 20 mins):
Age (per 5 years):
Baseline systolic BP (per 10mmHg):
Baseline diastolic BP (per 10mmHg):
Baseline total cholesterol (per 1 mmol/l):
Time index event to procedure (per 7days):
Variable
28/683 (4.1%)5/131 (3.8%)
10/220 (4.5%)17/404 (4.2%)6/190 (3.2%)
13/275 (4.7%)20/539 (3.7%)
2/72 (2.8%)31/749 (4.1%)
19/539 (3.5%)5/136 (3.7%)7/104 (6.7%)1/35 (2.9%)
15/360 (4.2%)1/22 (4.5%)14/292 (4.8%)3/134 (2.2%)
19/513 (3.7%)14/308 (4.5%)
28/718 (3.9%)5/103 (4.9%)
10/307 (3.3%)8/194 (4.1%)9/212 (4.2%)3/74 (4.1%)2/17 (11.8%)0/3 (0%)
19/427 (4.4%)14/394 (3.6%)
16/284 (5.6%)16/495 (3.2%)
24/609 (3.9%)
22/685 (3.2%)
33/821 (4%)
31/784 (4%)
31/785 (3.9%)
31/679 (4.6%)
32/818 (3.9%)
/ No. pts (%)No. events
1.000.93 (0.37,2.37)
1.000.93 (0.43,1.99)0.70 (0.26,1.88)
1.000.79 (0.40,1.55)
1.001.49 (0.36,6.10)
1.001.04 (0.40,2.74)1.91 (0.82,4.43)0.81 (0.11,5.88)
1.001.09 (0.15,7.89)1.15 (0.57,2.35)0.54 (0.16,1.83)
1.001.23 (0.62,2.41)
1.001.25 (0.49,3.15)
1.001.27 (0.51,3.15)1.30 (0.54,3.15)1.24 (0.35,4.40)3.61 (0.86,15.20)
1.000.80 (0.41,1.57)
1.000.57 (0.29,1.13)
0.95 (0.69,1.31)
0.89 (0.73,1.09)
1.08 (0.89,1.30)
1.13 (0.99,1.29)
1.30 (1.02,1.66)
0.96 (0.73,1.25)
1.01 (0.98,1.04)
ratio (95% CI)Risk
0.88
0.75
0.49
0.58
0.54
0.62
0.55
0.64
0.77
0.52
0.11
0.75
0.27
0.45
0.07
0.04
0.74
0.48
P-value
1.000.93 (0.37,2.37)
1.000.93 (0.43,1.99)0.70 (0.26,1.88)
1.000.79 (0.40,1.55)
1.001.49 (0.36,6.10)
1.001.04 (0.40,2.74)1.91 (0.82,4.43)0.81 (0.11,5.88)
1.001.09 (0.15,7.89)1.15 (0.57,2.35)0.54 (0.16,1.83)
1.001.23 (0.62,2.41)
1.001.25 (0.49,3.15)
1.001.27 (0.51,3.15)1.30 (0.54,3.15)1.24 (0.35,4.40)3.61 (0.86,15.20)
1.000.80 (0.41,1.57)
1.000.57 (0.29,1.13)
0.95 (0.69,1.31)
0.89 (0.73,1.09)
1.08 (0.89,1.30)
1.13 (0.99,1.29)
1.30 (1.02,1.66)
0.96 (0.73,1.25)
1.01 (0.98,1.04)
ratio (95% CI)Risk
0.88
0.75
0.49
0.58
0.54
0.62
0.55
0.64
0.77
0.52
0.11
0.75
0.27
0.45
0.07
0.04
0.74
0.48
P-value
1.05 .2 .5 1 2 4 6 10 15
Risk Ratio (95%CI)
63
In a multivariable model, only diastolic blood pressure remained a significant predictor
of the risk of stroke, MI or death within 30 days of the procedure in the presence of any
other included variables (RR 1.30 for each +10mmHg, 95% CI 1.02 to 1.66, p=0.04).
This model included only 785 patients for whom complete predictor data were
available.
Discussion and conclusion 14.4
14.4.1 Summary
The overall risk of stroke, MI or death within 30 days of CEA in ICSS was 4.0%. The
majority of these events were strokes, and the majority of events occurred before
discharge. Of the baseline demographic and vascular risk factors examined, only sex
and increasing diastolic blood pressure at baseline significantly predicted the risk of
stroke, MI or death. Female patients experienced roughly double the risk of male
patients of reaching the composite endpoint. In a multivariable model, only baseline
diastolic blood pressure remained a significant predictor of risk after adjustment for all
other included variables.
None of the surgical technical variables examined – type of anaesthesia, type of
arterial reconstruction, variation in medical therapy or shunt use – significantly
predicted risk.
14.4.2 Discussion
The finding that women undergoing CEA in ICSS experienced a higher risk of stroke,
MI or death than men is broadly consistent with findings from other randomized trials
and audits of carotid endarterectomy outcomes [138] [133] [139] [140], although recent
data from the Society for Vascular Surgery Vascular Registry suggest a similar
complication rate (around 4% risk of stroke, MI or death within 30 days of CEA) for both
men and women outside a clinical trial setting [141]. Although a statistically significant
difference was observed in ICSS this result was not confirmed in multivariable analysis.
This suggests that differences in baseline patient characteristics between male and
female patients might explain this difference in risk [118], although this analysis had
limited power to detect predictors because of the relative paucity of outcome events.
An alternative explanation for a higher observed risk of complications in women may be
that the smaller carotid artery diameter in women is associated with procedural stroke
due to in-situ thrombosis following more technically-demanding surgery [142] [143]
[139]. Or perhaps women, with an older age of onset of cerebrovascular disease than
64
men, might have more medical comorbidities that increase procedural risk [144]. In
addition, many of the post-procedural strokes experienced by ICSS patients
undergoing CEA occurred some days after the procedure, and it has been suggested
that women are at a higher risk of post-procedural embolism [145] [146].
Notably, women with symptomatic carotid stenosis may also experience less benefit in
terms of stroke prevention from CEA, particularly where the procedure is delayed [147].
Thus any risk-benefit discussion with female patients should highlight both a potential
increased risk of minor complications as well as a reduced overall benefit as compared
with men.
Time from the index event prompting randomization in ICSS to the date of surgery did
not appear to influence risk, but the majority of patients in ICSS waited more than 14
days after randomization for surgery [82] which exceeds the maximum recommended
delay in current guidelines [128]. In 2012, 33% of UK patients undergoing CEA were
operated on within 14 days of their symptoms. However, there is some evidence that
very early CEA, within 48 hours of symptoms, is associated with increased risk [148].
In ICSS surgical technique and mode of anaesthesia had less influence on short-term
outcome than patient characteristics. The largest randomized trial of general vs local
anaesthesia for CEA, the General Anaesthesia versus Local Anaesthesia for carotid
surgery trial (GALA), found no difference in the risk of stroke, MI or death within 30
days of surgery between the two groups [104] [103]. There is currently insufficient
evidence to suggest that routine shunt use reduces perioperative stroke or death [106],
by maintaining cerebral perfusion pressure in the presence of a tight contralateral
carotid stenosis, although there is some suggestion that patch closure of the artery, as
opposed to primary closure, is associated with a lower risk of perioperative ipsilateral
stroke [107].
The association between baseline blood pressure and outcome of surgery has also
been demonstrated previously in a systematic review including patients from the
European Carotid Surgery Trial [149]. Despite detailed baseline and follow-up
measurements, ICSS investigators did not collect perioperative blood pressures and
therefore it was not possible to assess the influence of perioperative haemodynamic
control on outcome. Patients who are transferred to Intensive Care or a similar
specialized unit after the procedure might be expected to have better optimization of
their blood pressures and medications. In ICSS 8.3% of patients undergoing CEA
experienced post-procedural hypertension which was associated with higher baseline
65
blood pressure [150] and in other studies has been shown to be associated with stroke
or death [151] [152]. What is not known is whether closer control of blood pressure
after stroke or TIA could reduce the risks of subsequent CEA. Although not statistically
significant, ICSS patients treated in centres with a policy of sending the patient to
intensive care afterwards experienced a much-reduced rate of complications, which
could in turn be due to closer monitoring of haemodynamic parameters and treatment
of excessively high or low blood pressures. Current guidelines for CEA have focussed
on patient selection, procedural technique and medication use [153]. Therefore given
our findings on diastolic blood pressure, and that more events occurred on the days
following the procedure than the day of the procedure itself, future work on the safety of
CEA should additionally address overall perioperative care and management of
physiological parameters.
14.4.3 Limitations
The risk of stroke, MI or death associated with carotid endarterectomy was acceptably
low in ICSS. However, a low number of outcome events limited the power to detect a
significant difference between subgroups of patients and limited the number of factors
supported in a multivariable model. Multiple statistical comparisons are made in this
analysis without correction, raising the possibility of a Type 1 (false positive) error.
Associations between baseline characteristics or processes of care and the outcome
do not necessarily imply causation, and this was not a randomized comparison
between surgical techniques.
Other factors or more generalized measures of health not recorded in ICSS, such as
the patient’s American Society of Anaesthesiology (ASA) grade, may be more useful
predictors of adverse outcome [154].
14.4.4 Conclusions
The relative safety of CEA in ICSS limited power to detect predictors of risk.
Nonetheless, diastolic blood pressure at baseline was a significant independent
predictor of the risk of stroke, MI or death within 30 days of carotid endarterectomy. It is
possible that careful attention to blood pressure control following stroke or TIA may
reduce the risks associated with subsequent CEA in patients with underlying carotid
stenosis, although concern exists about acute lowering of blood pressure following
stroke or TIA in the presence of carotid stenosis. The finding that around 20% of events
occurred on or after the day of discharge underlines the importance of post-discharge
66
follow-up to ensure that patients receive treatment for any late complications and to
ensure that surgical audit data include all late events. Surgical results should also be
carefully audited in female patients to ensure that the risk of complications remains
acceptably low.
67
15. Validation of existing risk scores for carotid
endarterectomy
Introduction 15.1
CEA remains the standard of care for patients with 50-99% symptomatic carotid
stenosis, as measured by the NASCET method [40]. However, despite improvements
in the care of surgical patients, and despite innovations in surgical and anaesthetic
technique for carotid endarterectomy, data from a systematic review of studies
published between 1994 and 2001 suggest that risk of stroke or death within 30 days of
surgery remained clinically significant at 6.5% (95% CI 4.3% to 8.7%) in studies
involving assessment of study endpoints by neurologists [96]. These studies are
comparable with ICSS in which 4.7% of patients randomized to CEA suffered stroke or
death [82] within 30 days of surgery.
Research since the landmark NASCET and ECST trials has attempted to define the
patient at high risk of complications following surgery. Complications of specific
concern in CEA include stroke, myocardial infarction, death, neck haematoma and
cranial nerve palsy. However, few risk factors or scoring systems found to be
significant predictors in one group of patients have been validated in other cohorts and
the usefulness of their application in everyday clinical decision-making remains
unclear. Risk scores may give patients and clinicians an individualized estimate of the
risk of suffering stroke or death within 30 days of endarterectomy, and could allow
calculation of the net benefit of surgery in long-term stroke prevention.
The performance of three risk scores in patients allocated to surgery the International
Carotid Stenting Study is assessed in this chapter. The patient populations from which
these scores derived most closely match the inclusion criteria for ICSS and comprise
clinical characteristics that are apparent at baseline visit.
Rothwell et al. [98] developed a prognostic model using data from patients undergoing
CEA in the European Carotid Surgery Trial (ECST) in the 1990s. ECST enrolled
patients with symptomatic carotid stenosis and randomized them to either carotid
endarterectomy or medical treatment [36]. A model to predict the risk of stroke or death
within 30 days of the procedure was calculated in patients with 0-69% stenosis of the
symptomatic artery and later validated in patients with 70-99% stenosis. Predictors of
risk were female sex, peripheral vascular disease, and high systolic blood pressure.
68
Tu et al. [155] developed a score to predict the risk of stroke or death within 30 days of
CEA in 6038 patients undergoing surgery in Ontario, Canada. 69% of those patients
were symptomatic, and 95% had ICA stenosis between 50-100%. The average age
was 68 years. Significant predictors of stroke were TIA or stroke in the 6 months prior
to surgery, atrial fibrillation, heart failure, diabetes and contralateral carotid occlusion.
Finally, the prediction model developed by Kuhan et al. [156] examined the patient and
procedural characteristics of 839 patients undergoing CEA in 2 vascular units in the
United Kingdom between 1992 and 1999. 87% of these patients were symptomatic,
60% were male and the median age was 68. Significant risk factors for stroke or death
within 30 days of CEA were “heart disease” (comprising angina, myocardial infarction,
heart failure or arrhythmia), diabetes and stroke as the index event prompting
intervention. None of the surgical variables included in the study (shunt use, patch use,
operating surgeon, vascular unit or the side of operation) significantly predicted risk.
A number of other publications address risk factors for complications of CEA, but were
not validated in this patient cohort as they included factors not measured at baseline in
ICSS patients [157] [158] [159] [160], included a large proportion of asymptomatic
patients [158] [161] [162] or focus on operator, hospital or process of care factors that
are modifiable and therefore of limited use in predicting the risk in individual patients
[163] [164] [165].
Methods 15.2
The performance of the three risk scores predicting the risk of stroke or death within 30
days of CEA was evaluated [98] [155] [156]. To test the performance of each risk
score, patients were assigned “risk points” according to their baseline characteristics.
The characteristics predicting risk in each risk score are outlined in Tables 3 to 5
below. Baseline patient characteristics were collected by investigators in ICSS at the
time of the patient’s enrolment into the trial and sent to the central trial office in London.
The occurrence of stroke or death within 30 days of the procedure was examined in a
per-protocol population of patients in ICSS in whom the allocated procedure was
initiated. CEA was deemed to have been initiated if local or general anaesthesia was
administered prior to the planned commencement of surgery. The adjudication of
events is described in more detail in Chapters 13 and 14. In brief, stroke was defined
as “an acute disturbance of focal neurological function lasting more than 24 hours
resulting from intracranial vascular disturbance”, and confirmatory evidence was
69
required where available in the form of CT or MR brain imaging. Centres reporting
death within 30 days of the procedure provided either a death certificate or autopsy
report if available.
70
Table 3. Allocation of risk points in the Rothwell risk score
Study Risk factors Risk points
Rothwell et al.
(1999) [98]
Female sex
1
Peripheral vascular disease
1
Systolic BP > 180mmHg 1
71
Table 4. Allocation of risk points in the Tu risk score
Study Risk factors Risk points
Tu et al.
(2003) [155]
TIA or stroke in the 6 months prior to surgery 1
Atrial fibrillation 1
Heart failure 1
Diabetes 1
Contralateral carotid occlusion 1
72
Table 5. Allocation of risk points in the Kuhan risk score
Study Risk factors Risk points
Kuhan et al.
(2001)
[156]
Heart disease (any of angina, MI, heart failure, arrhythmia) 1
Stroke as the index event prompting surgery 1
Diabetes (any type) 1
73
A 95% confidence interval for the risk of stroke or death within 30 days of the
procedure was generated, and the performance of the risk score across groups with
increasing score was tested by means of a chi-squared (χ2) test for trend. The
“predicted” risk of stroke or death within 30 days of surgery represents the result of
applying the risk score to the population in which it was originally developed. Groups in
which no events occurred were combined with lower-risk groups to enable the
statistical comparison to be carried out. Receiver operator curves (ROC) were plotted
to assess the performance of each risk score, and the area under the curve (AUC) was
calculated with 95% confidence intervals and significance test. Statistical analyses
were performed using GraphPad (GraphPad Software, 2013. GraphPad Instat version
3.00 for Windows 95. San Diego, California: www.graphpad.com).
Results 15.3
The baseline characteristics of patients included in this analysis, and how many
patients possessed characteristics which feature in each of the three risk scores, are
presented in detail in Chapter 14.3.2 and Figure 7. 70.4% of patients were male, and
52% were aged 70 years or more. 59/821 (7.2%) patients had systolic blood pressure
>180mmHg, and 236 patients (28.7%) had “heart disease” as defined in the Kuhan risk
score (any of angina, MI, heart failure or arrhythmia). 174 patients (21.2%) were
diabetic, and 131 (16.0%) had a diagnosis of peripheral vascular disease.
28/821 (3.4%) patients allocated to CEA in ICSS, in whom the procedure was initiated,
suffered stroke or death within 30 days of the procedure.
15.3.1 Rothwell (1999) risk score
The results of applying the Rothwell score to ICSS patients allocated to CEA are
detailed in Table 6. The risk of stroke or death within 30 days of surgery was 2.5%
(95% CI 1.3% to 4.4%) in those with a score of 0, 4.4% (95% CI 2.6% to 7.4%) in those
with a score of 1 and 5.4% (95% CI 1.3% to 15.2%) in those with a score of 2. 2
patients had a score of 3 but neither suffered an event. There was an increase in the
observed risk of an outcome event with increasing score that was not statistically
significant (p=0.10). 25/28 (89.2%) events occurred in the lowest risk groups (score 0
to 1).
A receiver operator curve (ROC) was plotted for the results of applying the Rothwell
score to the ICSS dataset and is presented in Figure 8. The area under the curve was
0.58 (95% CI 0.48 to 0.69, p=0.14), indicating poor performance.
74
Table 6. Results of application of the Rothwell (1999) risk score to 821 patients undergoing CEA in ICSS
Score Number at risk Number of events Predicted event rate % (95% CI) Observed event rate % 95% CI for observed event rate
0 448 11 4.7 (2.6 to 7.6) 2.5 (1.3 to 4.4)
1 315 14 7.3 (4.3 to 11.0) 4.4 (2.6 to 7.4)
2 56 3 12.1 (5.0 to 23.0) 5.4 (1.3 to 15.2)
3 2 0 16.7 (4.2 to 64.0) 0.0
Totals 821 28
χ2 test for trend p=0.10
75
Figure 8. Receiver operator curve for application of the Rothwell (1999) risk score
to 821 patients undergoing CEA in ICSS
Darker line represents receiver operator curve
Straight line represents area under curve =0.5
76
15.3.2 Tu (2003) risk score
The results of applying the Tu score to the ICSS dataset are detailed in Table 7. The
risk of stroke or death within 30 days of the procedure was 2.6% (95% CI 0.6% to
7.6%) in those with a score of 0, 3.4% (95% CI 2.0% to 5.4%) in those with a score of 1
and 4.7% (95% CI 2.4% to 8.9%) in those with a score of 2. Patients with higher scores
did not experience an event. The increase in the risk of an event with increasing score
did not reach statistical significance (p=0.49). All events occurred in lower risk groups
(score 0 to 2).
In a receiver operator curve plot of these results, presented in Figure 9, the AUC was
0.526 (95% CI 0.42 to 0.63, p=0.64) indicating poor performance.
77
Table 7. Results of application of the Tu (2003) risk score to 821 patients undergoing CEA in ICSS
Score Number at risk Number of events Predicted event rate % (95% CI) Observed event rate % 95% CI for observed event rate
0 117 3 3.3 2.6 (0.6 to 7.6)
1 477 16 6.1 3.4 (2.0 to 5.4)
2 190 9 9.5 4.7 (2.4 to 8.9)
3 33 0 9.8 0.0 N/A
4 4 0 15.8 0.0 N/A
5 0 0 N/A N/A N/A
Totals 821 28
χ2 test for trend p=0.49
78
Figure 9. Receiver operator curve for application of the Tu (2003) risk score to
821 patients undergoing CEA in ICSS
Darker line represents receiver operator curve
Straight line represents area under curve =0.5
79
15.3.3 Kuhan (2001) risk score
The results of applying the Kuhan score to the ICSS dataset are outlined in Table 8
below. The risk of stroke or death within 30 days of CEA was 2.2% (95% CI 0.9% to
4.8%) in patients with a score of 0, 3.7% (95% CI 2.1% to 6.3%) in those with a score
of 1, 4.2% (95% CI 1.9% to 8.6%) in those with a score of 2 and 5.9% (95% CI 0.7% to
20.0%) in those with a score of 3. 19/28 (67.9%) events occurred in lower risk groups
(scores 0 to 1).
The area under the receiver operator curve, illustrated in Figure 10, was 0.58 (95% CI
0.47 to 0.68, p=0.18), indicating poor performance.
80
Table 8. Results of application of the Kuhan (2001) risk score to 821 patients undergoing CEA in ICSS
Score Number at risk Number of events Predicted event rate % (95% CI) Observed event rate % 95% CI for observed event rate
0 272 6 1.4 (0.9 to 2.1) 2.2 (0.9 to 4.8)
1 349 13 3.2 (2.1 to 4.8) 3.7 (2.1 to 6.3)
2 166 7 7.2 (2.5 to 19.0) 4.2 (1.9 to 8.6)
3 34 2 15.3 (2.8 to 53.1) 5.9 (0.7 to 20.0)
Totals 821 28
χ2 test for trend p=0.15
81
Figure 10. Receiver operator curve for application of the Kuhan (2001) risk score to
821 patients undergoing CEA in ICSS
Darker line represents receiver operator curve
Straight line represents area under curve =0.5
82
Discussion and conclusion 15.4
15.4.1 Summary
Each of the three risk scores tested in this analysis demonstrated a similar pattern of
results: a small increase in the risk of stroke or death within 30 days of the procedure
as the score increased, a lower observed risk of an event in ICSS than predicted, and
the majority of events occurring in those groups of patients predicted to be at lower
risk. None of the scores tested were sufficiently discriminating to be able to
demonstrate a statistically significant difference in risk between groups of patients with
different scores and all showed poor sensitivity and specificity.
15.4.2 Limitations
The ability to accurately test many existing published risk scores was limited by the
data collected at baseline in ICSS, and the type of patients enrolled in the study. The
ability to detect significant differences in risk between groups predicted to be at low or
high risk was limited by the relative safety of CEA in ICSS and therefore the low
number of outcome events.
15.4.3 Discussion
“Complications” following any surgical procedure are common throughout the world,
occurring in up to 7% of those undergoing surgery even after specific interventions
designed to reduce the risk of adverse outcome [93]. CEA may be different to some
other types of major surgery due to the higher risk of stroke. Therefore there is a need
for unique risk scores to identify the patient at high risk during revascularization. To be
useful to clinicians, clinical risk scores should be effective, accurate and generalizable
to other patient populations [166] [167].
The risk models analysed in this chapter did not satisfactorily discriminate between
high- and low-risk groups of patients, and indeed excluding patients in higher risk
groups for undergoing surgery would not have prevented the majority of periprocedural
events.
Other reasons that these models may not have performed so well include differences
between the inclusion and exclusion criteria in the studies from which they are derived
as compared to the criteria in ICSS, the influence of unmeasured factors such as
operator experience or case volume [168], the type of hospital the surgery is carried
83
out in [165], and the influence of vascular features not routinely recorded at baseline in
ICSS, such as ulcerated plaque or fresh thrombus in the carotid artery, which may
double the risk of stroke after CEA [158]. There is also a need to update risk models as
the higher risks observed in original trials as compared with new trials means that the
models tested all over-estimate risk, a common finding in prediction models [169].
Other authors have examined the implications of applying the Rothwell score to a
population of patients with symptomatic carotid stenosis in a single unit. Their results
show a similar distribution of risk scores, with patients at lower predicted surgical risk
(score 0 or 1) making up the majority of those undergoing surgery [170]. However, the
rate of major stroke or death within 31 days of surgery in this study was only 0.75%,
and the authors conclude that excluding high-risk patients from surgery in this series
would have exposed the patients to a high risk of stroke on medical therapy alone with
no reduction in the risk of a major event in the group undergoing surgery.
As demonstrated in the European Carotid Surgery Trial, the factors that determine of
the risk of surgery may well be different from those that determine the risk of stroke
without treatment [98], and therefore patients should not be excluded from surgery
because of a high predicted risk of CEA without taking into consideration the risk of
stroke on secondary prevention medications alone [169]. In addition, the decision to
operate or not is made more complex by the availability of CAS, which may in turn
have different risk factors for periprocedural stroke or death as discussed in Chapter
17. Authors drawing on data from the European trials of stenting vs surgery for
symptomatic carotid stenosis have developed a clinical rule based on the patient’s sex,
age, contralateral carotid occlusion and the presence of restenosis (the “SCAR” rule
[171]) to address this issue and enable practitioners to choose between the two
revascularization procedures in patients where revascularization is the preferred
treatment option.
15.4.4 Conclusions
The observed risk of CEA in ICSS was acceptably low, and only 3.4% of patients
suffered stroke or death within 30 days of the procedure. In each of the three risk
scores evaluated, increasing score was associated with an increasing observed risk of
the outcome, although this association did not reach statistical significance in any of
the scores tested. The ability of the scores to discriminate between patients at low and
high risk was poor as demonstrated by the area under the receiver operator curve
measurements. Most events in ICSS occurred in groups of patients predicted to be at
84
low risk, and thus the overall performance of these scores was unsatisfactory, and
routine use of the scores to predict operative risk is not supported.
Patient risk factor profiles and the nature of surgical care for endarterectomy have
changed since the landmark NASCET and ECST trials, and influence of risk factors for
surgical complications from older studies may have diminished. The availability of non-
invasive detailed plaque imaging raises the possibility of developing modern and
better-performing risk scores for the procedural risk of CEA or the risk of stroke on
medical therapy alone, drawing on data from large-scale studies or ongoing trials
randomizing patients to CEA such as ECST-2 [172], SPACE-2 [173] and ACST-2 [174].
85
16. Risk factors for, and impact of, cranial nerve palsy and
haematoma following carotid endarterectomy
Introduction 16.1
The results of landmark trials of carotid surgery versus medical therapy, NASCET
[175], ECST [36] and the Veterans’ Affairs Trial [39] provided evidence that carotid
endarterectomy for symptomatic atherosclerotic carotid stenosis greater than 50%
provided a better long-term reduction in the risk of recurrent ipsilateral stroke than
medical therapy. Subsequent trials of CEA vs carotid stenting have suggested that, in
the short-term at least, surgery remains the safest treatment for carotid stenosis, with
CAS carrying a higher risk of any stroke, myocardial infarction or death within 30 days
of the procedure in meta-analysis (OR 1.44, 95% CI 1.15 to 1.80) [176]. However, the
primary endpoints of these trials and analyses do not contain outcomes that specifically
relate to the additional hazards of a surgical incision in the neck – namely cranial nerve
palsy (CNP) and haematoma.
Carotid stenting developed as an alternative to CEA in part to avoid these hazards of a
surgical incision. The higher procedural risk associated with CAS in ICSS was mainly
due to an excess of non-disabling stroke [82], and this disadvantage could be balanced
by an increased incidence of cranial nerve palsy in the surgical group.
CNP and haematoma are less extensively studied, but have long been recognised as
complications of surgery [47], and these “minor outcomes” have been associated with
an increased risk of major outcomes following surgery such as stroke or death [177].
Haematoma has been associated with an increased time spent in critical care following
surgery [178]. Published case series have highlighted frequently-affected cranial
nerves which include facial, vagal, hypoglossal and accessory [47] [48] [49] [50], and
there is a potential for these injuries to cause significant postoperative disability.
In this chapter the incidence and severity of CNP and haematoma in ICSS are
examined to determine whether these complications merit consideration in the
selection of revascularization procedure. Risk factors for their development are
analysed.
86
Methods 16.2
The ICSS trial protocol is summarized in Chapter 13 and presented in full in Appendix
IV. As reported in Chapter 14, the type of anaesthetic or surgical technique for CEA
were not specified in the trial protocol, but all patients were required to receive medical
therapy for secondary prevention of stroke as appropriate. A technical case report form
was completed for each endarterectomy procedure in ICSS and the occurrence of CNP
or haematoma was reported by investigators. CNP was defined as “weakness or
sensory impairment in the distribution of one of the cranial nerves attributed to
treatment”. Each patient was re-assessed at their one-month follow-up appointment by
a neurologist or investigator with an interest in stroke. CNPs were adjudicated at the
trial office based on the best available clinical evidence, and deemed “disabling” if the
patient’s modified Rankin Score (mRS) increased to 3 or more at 30 days post-
procedure where that increase was thought due to CNP. A questionnaire requesting
additional information about the clinical impact of an event and whether or not
symptoms eventually resolved (Appendix II) was sent to centres reporting CNP.
Haematoma was defined as “bleeding attributed to the treatment of carotid narrowing”.
Severe haematoma was defined as a haematoma “requiring new surgery, transfusion
or prolonging hospital stay”.
16.2.1 Statistical methods
The data were analysed per-protocol, including only those patients in ICSS in whom
the allocated CEA procedure was initiated. A procedure was deemed to have been
initiated if local or general anaesthesia was administered prior to commencement of
surgery. Patients who crossed-over to CAS or medical treatment were excluded. Risk
factors for CNP and haematoma were examined sequentially in a univariable binomial
regression analysis using maximum likelihood estimation. The risk ratio for each factor
is given with corresponding 95% confidence interval. Wald tests were used for
continuous and binary predictors, with an overall likelihood ratio test for categorical
predictors of more than two levels. A multivariable model was developed using a
forward stepwise-based approach with a significance level of 0.1 for predictors in
univariable analysis accepted for inclusion in the multivariable model. Patients with
missing data were excluded from each relevant analysis. Analyses were performed
using Stata (Stata Corp. 2011. Stata Statistical Software: Release 12. College Station,
TX, USA: Stata Corp LP).
87
Results – cranial nerve palsy 16.3
858 patients in ICSS were randomized to CEA, of whom 821 had their allocated
procedure initiated (95.7%). 4/821 patients (0.5%) died after initiation of the procedure
but before 30-day follow-up. 45/821 (5.5%) patients were reported to have developed
CNP between initiation of the procedure and 30-day follow-up. The results of an
adjudication of which cranial nerves are affected are presented in Table 9. The facial
nerve was involved in 23 patients, vagus nerve in 6, hypoglossal in 13,
glossopharyngeal in 4, accessory in 1 and trigeminal in 1. The nerve affected was
undetermined in two patients. In five patients more than one cranial nerve was
affected. It was not possible to determine the nerve affected in two patients.
One patient was judged to have disabling CNP at 30-day follow-up due to
glossopharyngeal nerve palsy. This caused impairment of swallowing with subsequent
need for naso-gastric feeding.
In those patients in whom CNP resolution was confirmed (n=20), the median estimated
duration of symptoms was 30 days (minimum 2 days, maximum 520 days). Permanent
injury (i.e. without resolution of symptoms by the end of their follow-up in the trial) was
reported in two patients who were followed-up over 6.4 and 3.1 years respectively.
88
Table 9. Summary of cranial nerve palsies following endarterectomy in ICSS
Cranial nerve Number of
postoperative
CNPs (n=50 in 45
patients)
Number of
disabling
CNPs (mRs≥3
at 1 month
due to CNP)
Number of
CNPs confirmed
persisting after
30 days
Facial 23 0 4
Hypoglossal 13 0 2
Vagus 6 0 4
Accessory 1 0 0
Glossopharyngeal 4 1 1
Trigeminal 1 0 0
Undetermined 2 0 0
89
The results of univariable regression analysis to determine risk factors for the
development of CNP within 30 days of the procedure are presented in Figure 11 below.
The risk of CNP was increased in female patients (RR 1.90, 95% CI 1.08 to 3.36,
p=0.03) and decreased in those with a high degree of contralateral carotid artery
stenosis. Other demographic factors did not predict the risk of CNP. Technical factors,
including the type of arterial reconstruction (standard closure, patch closure, eversion
endarterectomy or vein interposition), type of anaesthesia or shunt use, were not
significant predictors of risk.
90
Figure 11. Univariable predictors of the risk of cranial nerve palsy within 30 days of
endarterectomy in 821 ICSS per-protocol participants in whom the procedure
was initiated
91
Figure 11 continued
92
Figure 11 continued
93
The result of multivariable analysis of independent predictors of the risk of CNP is
given in one possible model in Table 10 below. In this model, independent predictors of
risk were cardiac failure (RR 2.66, 95% CI 1.11 to 6.40, p=0.03), female sex (RR 1.80,
95% CI 1.02 to 3.20, p=0.04), the degree of contralateral carotid artery stenosis and a
time of greater than 14 days from randomization in the trial to surgery (RR 3.33, 95%
CI 1.05 to 10.57, p=0.04).
The risk of cranial nerve palsy in the stenting group in ICSS was 1/828 (0.12%), and on
adjudication this complication was attributed to a second procedure (CEA) carried out
before 30-day follow-up.
94
Table 10. Independent predictors of risk of cranial nerve palsy within 30 days of
endarterectomy in 805* ICSS per-protocol participants in whom the procedure
was initiated
Variable Adjusted Risk Ratio (95% CI) Adjusted p-value
Cardiac failure 2.66 (1.11 to 6.40) p=0.03
Female sex 1.80 (1.02 to 3.20) p=0.04
Time from randomization
to treatment >14 days
3.33 (1.05 to 10.57) p=0.04
Degree of
contralateral stenosis
0-50%
50-69%
>70%
1.00
1.18 (0.62 to 2.27)
0.13 (0.02 to 0.91)
Overall p<0.01
p=0.62
p=0.04
* patients with missing data excluded from this analysis
95
Results – haematoma 16.4
50/828 (6.1%) patients in the CEA arm of ICSS who had their procedure initiated
developed haematoma. 28/828 (3.4%) patients had severe haematoma, requiring
transfusion, re-operation or prolonging hospital stay. 12 of those 50 patients who
suffered haematoma also developed CNP, and there was a significant association
between these complications (p<0.01, Fisher’s exact test).
The results of univariable analysis of factors predicting the risk of haematoma within 30
days of CEA are presented in Table 11 below. Those factors that increased the risk of
haematoma were preoperative prescription of anticoagulant medication (RR 1.83, 95%
CI 1.04 to 3.23, p=0.04), atrial fibrillation (RR 2.29, 95% CI 1.08 to 4.85, p=0.03),
previous cardiac bypass graft surgery (RR 2.46, 95% CI 1.37 to 4.42, p<0.01) and the
duration of arterial clamping (RR per each 20 minutes 1.13, 95% CI 1.04 to 1.24,
p<0.01). Factors that decreased the risk of haematoma were shunt use (RR 0.54, 95%
CI 0.29 to 0.99, p=0.05), the prescription of an antiplatelet agent pre-procedure (RR
0.44, 95% CI 0.21 to 0.93, p=0.03) and each increase in 1mmol/l of cholesterol at
baseline (RR 0.69, 95% CI 0.55 to 0.88, p<0.01). Other demographic and technical
factors were not statistically significant predictors of the risk of haematoma.
96
Table 11. Univariable predictors of risk of haematoma within 30 days of endarterectomy in 821 ICSS per-protocol participants in whom the
procedure was initiated
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Type of reconstruction Standard 9.3 1 0.26
Eversion 8.2 0.87 (0.31 to 2.48)
Patch 5.2 0.56 (0.31 to 1.02)
Vein interposition 0 N/A
Shunt use No 7.5 1 0.05
Yes 4.0 0.54 (0.29 to 0.99)
97
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Type of anaesthesia GA or LA/GA combined 6.5 1 0.69
LA alone 5.6 0.86 (0.41 to 1.79)
Antiplatelet agent pre-procedure No 13.5 1 0.03
Yes 5.9 0.44 (0.21 to 0.93)
Two or more antiplatelet agents pre-procedure No 6.0 1 0.50
Yes 7.3 1.21 (0.69 to 2.11)
98
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Anticoagulant pre-procedure No 5.5 1 0.04
Yes 10.1 1.83 (1.04 to 3.23)
Antiplatelet agent post-procedure No 9.2 1 0.26
Yes 6.1 0.66 (0.32 to 1.36)
Anticoagulant post-procedure No 5.6 1 0.11
Yes 8.8 1.58 (0.91 to 2.76)
99
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Sex Male 5.2 1 0.10
Female 8.2 1.59 (0.92 to 2.74)
Age <70 years 6.1 1 0.96
≥70 years 6.0 0.99 (0.58 to 1.69)
Treated hypertension No 5.0 1 0.36
Yes 6.6 1.34 (0.71 to 2.52)
100
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Cardiac failure No 6.1 1 0.85
Yes 6.8 1.12 (0.36 to 3.45)
Angina No 6.1 1 0.77
Yes 6.9 1.15 (0.47 to 2.79)
Previous MI No 5.4 1 0.07
Yes 9.3 1.72 (0.95 to 3.11)
101
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Previous CABG No 5.1 1 <0.01
Yes 12.6 2.46 (1.37 to 4.42)
Atrial fibrillation No 5.7 1 0.03
Yes 13.0 2.29 (1.08 to 4.85)
Other cardioembolic source of embolus No 6.1 1 0.99
Yes 6.3 1.02 (0.15 to 6.92)
102
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Diabetes No 6.3 1 0.81
Yes 5.7 0.92 (0.47 to 1.80)
Peripheral arterial disease No 6.3 1 0.68
Yes 5.3 0.85 (0.39 to 1.85)
Smoking status Never smoked 4.1 1 0.23
Former smoker 7.4 1.82 (0.88 to 3.75)
Current smoker 5.8 1.42 (0.60 to 3.34)
103
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Treated hyperlipidaemia No 6.5 1 0.73
Yes 5.9 0.91 (0.52 to 1.59)
Degree of stenosis in treated artery 50 – 69% 6.9 1 0.75
70 – 99% 6.0 0.87 (0.36 to 2.11)
Degree of stenosis in contralateral artery 0 to 49% 6.1 1 0.39
50 to 69% 7.4 1.20 (0.61 to 2.38)
70 to 99% 2.9 0.47 (0.15 to 1.51)
Occluded 8.6 1.40 (0.45 to 4.34)
104
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Nature of ipsilateral index event Stroke 5 1 0.32
Retinal stroke 4.5 0.91 (0.13 to 6.51)
TIA 6.2 1.23 (0.65 to 2.33)
Amaurosis fugax 9.7 1.94 (0.98 to 3.85)
Multiple ipsilateral events prior to randomization No 5.5 1 0.33
Yes 7.1 1.31 (0.76 to 2.25)
Prior ipsilateral stroke No 5.8 1 0.45
Yes 7.8 1.33 (0.64 to 2.75)
105
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Baseline Rankin score 0 6.5 1 0.42
1 6.2 0.95 (0.48 to 1.90)
2 7.1 1.09 (0.57 to 2.07)
3 1.4 0.21 (0.03 to 1.52)
4 5.9 0.90 (0.13 to 6.33)
5 0 N/A
Side of procedure Left 5.9 1 0.77
Right 6.3 1.08 (0.63 to 1.86)
106
Table 11 continued
Variable Category 30-day risk of haematoma (%) Risk Ratio (95% CI) P-value
Time from randomization to treatment ≤14 days 4.6 1 0.41
>14 days 6.4 1.39 (0.64 to 3.03)
Clamping time (Per 20 minutes of clamp applied) 6.9 1.13 (1.04 to 1.24) <0.01
Duration of surgery (Per 20 minutes of operation length) 6.4 0.97 (0.86 to 1.10) 0.66
Age (Per 5 years of age) 6.1 1.02 (0.88 to 1.18) 0.84
Baseline systolic blood pressure (Per 10mmHg of blood pressure) 6.3 1.08 (0.98 to 1.21) 0.13
Baseline diastolic blood pressure (Per 10mmHg of blood pressure) 6.2 0.92 (0.74 to 1.14) 0.45
Baseline cholesterol (Per each mmol/l total cholesterol) 6.3 0.69 (0.55 to 0.88) <0.01
107
The results of multivariable analysis of predictors of the risk of haematoma are
presented in Table 12 which illustrates one possible model. Independent predictors of
increased risk were female sex (RR 2.03, 95% CI 1.13 to 3.62, p=0.02), atrial fibrillation
(RR 2.38, 95% CI 1.07 to 5.27, p=0.03) and the prescription of an anticoagulant drug
pre-procedure (RR 1.86, 95% CI 1.01 to 3.42, p=0.05). Independent factors that
decreased the risk were shunt use (RR 0.40, 95% CI 0.21 to 0.80, p<0.01) and
baseline cholesterol level (RR 0.68 per 1mmol/l increase, 95% CI 0.54 to 0.86,
p<0.01).
108
Table 12. Independent predictors of the risk of haematoma within 30 days of endarterectomy in 639 ICSS per-protocol participants in whom the
procedure was initiated
Variable Risk Ratio (95% CI) Adjusted p-value
Anticoagulant pre-procedure 1.86 (1.01 to 3.42) 0.05
Shunt use 0.40 (0.21 to 0.80) <0.01
Cholesterol (per each mmol/l) 0.68 (0.54 to 0.86) <0.01
Female 2.03 (1.13 to 3.62) 0.02
Atrial fibrillation 2.38 (1.07 to 5.27) 0.03
* patients with missing data excluded from this analysis
109
Results – impact of adding CNP to ICSS trial primary outcomes 16.5
Table 13 below details the impact of adding cranial nerve palsy to the primary and
secondary outcomes of ICSS in interim analysis [82]. For the combined outcome of
stroke, MI, death or cranial nerve palsy within 30 days of CEA there was no significant
difference in risk between CAS and CEA groups (RR 0.81, 95% CI 0.59 to 1.12,
p=0.20). For the combined outcome of disabling stroke, disabling CNP or death only
there was also no significant difference in risk (RR 1.41, 95% CI 0.79 to 2.51, p=0.24).
110
Table 13. Composite outcome events within 30 days of CEA versus CAS in ICSS
Endpoint
(within 30 days of the procedure)
CAS (n=828)
Events (%)
CEA (n=821)
Events (%)
Risk ratio
(95% CI)
Risk difference
(95% CI)
P-value
(Chi-squared)
Stroke, MI or death 61 (7.4%) 33 (4.0%) 1.83 (1.21, 2.77) 3.3% (1.1, 5.6) <0.01
Stroke, MI, death or CNP 62 (7.5%) 76 (9.3%) 0.81 (0.59, 1.12) -1.8% (-4.4, 0.9) 0.20
Disabling stroke or death 26 (3.1%) 18 (2.2%) 1.43 (0.79, 2.59) 0.9% (-0.6, 2.5) 0.23
Disabling stroke, disabling CNP or
death
27 (3.3%) 19 (2.3%) 1.41 (0.79, 2.51)
0.9% (-0.6, 2.5) 0.24
111
Results – influence of sex on combined outcomes 16.6
Table 14 below summarises the findings of both Chapter 14 and this chapter
regarding the influence of sex on the outcomes of CNP, haematoma, or stroke,
MI or death. Women had a statistically significantly higher risk of each of these
outcomes in the surgical arm of ICSS. When these outcomes were combined, the
risk of any surgical complication (i.e. CNP, haematoma, stroke, MI or death) in
the endarterectomy arm of ICSS was considerably higher in women (19.7%)
compared with men (10.9%, p<0.01).
112
Table 14. Results of univariable binomial regression analyses of female vs male in
821 per-protocol CEA patients in whom the allocated procedure was initiated
30-day outcome Number events / Number patients (%) Risk Ratio (95%
CI), p-value
Females Males
CNP 20/243 (8.2%) 25/578 (4.3%) 1.90 (1.08, 3.36),
p=0.03
Haematoma 20/243 (8.2%) 30/578 (5.2%) 1.59 (0.92, 2.74),
p=0.10
Stroke, MI or death 15/243 (6.2%) 18/578 (3.1%) 1.98 (1.02, 3.87),
p=0.05
CNP, haematoma,
stroke, MI or death
48/243 (19.7%) 63/578 (10.9%) 1.81 (1.28, 2.56),
p<0.01
113
Discussion and conclusion 16.7
16.7.1 Summary
5.5% of ICSS patients who were allocated to carotid endarterectomy, and in whom the
procedure was initiated, developed CNP. 6.1% developed haematoma within 30 days.
There was a statistically significant association between the two outcomes. The most
commonly-affected nerves were facial and hypoglossal, although only one CNP was
classified as disabling at 30 days. Independent predictors that increased of the risk of
cranial nerve palsy were cardiac failure, female sex, the degree of contralateral carotid
stenosis and a time between randomization and surgery of >14 days. The risk of
haematoma was raised in female patients, those with atrial fibrillation and those taking
anticoagulant medications pre-procedure.
16.7.2 Discussion - cranial nerve palsy
The European Carotid Surgery Trial found a motor CNP risk of 5.1% and a long-term
CNP risk of 0.5% at 4 months [179]. Similarly, the North American Symptomatic
Carotid Endarterectomy Trial found a postoperative risk of CNP of 8.6%, the majority of
which were mild in severity [175]. The finding of a 5.5% risk of CNP in ICSS, the 5.6%
risk of CNP in 6878 patients included in the Vascular Study Group of New England
study [180] and recent results from CREST showing a CNP risk of 4.6%, four-fifths of
which resolved within 1 year [181], suggests that the risk of CNP following CEA has
remained relatively constant over time. However, reassuringly, the risk of disabling
CNP is small – around 1 in 1000 operations. The symptoms of CNP may persist for
several weeks, but half of patients in whom the lesion resolves will experience an
improvement within 30 days.
The higher risk in female patients has not previously been described, and this finding is
worth confirming in other patient groups. It may be that surgical anatomy is more
challenging in female patients, who have a smaller diameter of carotid artery on
average [142], leading to inadvertent trauma to the cranial nerves running adjacent.
Female patients were also at risk of haematoma, perhaps for the same reason.
16.7.3 Discussion - haematoma
Female sex, anticoagulants and atrial fibrillation were independent risk factors for the
development of neck haematoma in this patient population. Although less well-studied
than cranial nerve palsy and major outcomes, previously-identified causes of
114
haematoma following CEA have included postoperative hypertension, anticoagulants
and antiplatelet agents [182] [183] (which might cause increased bleeding), as well as
inadequate surgical drainage [184]. Independent risk factors in another study were
found to be non-reversal of heparin, intraoperative hypotension and carotid shunt
placement [178].
A risk of haematoma in ICSS of 6.1% is similar to the risk of “wound complications”
(haematoma and infection) of 6.0% seen in a large multicentre case series of 1998
patients undergoing CEA [51], and a severe haematoma risk of 3.4% is similar to the
reported risk of requiring re-exploration in patients undergoing CEA while taking
antiplatelet medication of 3.6% [52].
Antiplatelet agents prevent recurrent stroke in patients with symptomatic carotid
stenosis [21] [185] [186] [187]. However, there is concern that clopidogrel, or
clopidogrel in combination with aspirin increases the risk of haematoma following
surgery [52]. Importantly, in ICSS there was no evidence that dual antiplatelet therapy
increased the risk of haematoma. Indeed, the clinical practise of most UK surgeons is
to continue single or dual antiplatelet therapy in the perioperative period where patients
are already taking these medications [188].
The association between neck haematoma and preoperative anticoagulation is not
unexpected, but our results reinforce the importance of careful attention to haemostasis
before closure of the wound.
16.7.4 Discussion – adding CNP to composite trial outcomes
Some systematic reviews have included CNP in a composite outcome of death or
neurological complications up to 30 days after treatment [176]. In ICSS, the total
number of events in the composite outcome of any stroke, myocardial infarction or
death was greater after CEA compared with CAS, the numbers in the composite
outcome of disabling stroke, disabling CNP or death were greater after CAS compared
with CEA, but neither difference was statistically significant. Thus it is unlikely to be
useful to include CNP in the composite endpoints of future trials, and it would not be
appropriate exclude patients from CEA because of an increased risk of CNP as
compared with CAS.
115
16.7.5 Discussion – the influence of sex on composite trial outcomes
Women were at higher risk of any surgical complication within 30 days of CEA in ICSS.
Indeed, the combined risk of CNP, haematoma, stroke, MI or death reached nearly
20%. While this finding may inform discussions with patients about procedural risk, it
should be remembered that female sex was not an independent predictor of the risk of
a major complication (stroke, MI or death) and that the impact of CNP appears to be
limited in duration and severity. Nonetheless, outcome data captured in clinical trials
may underestimate the quality of life impact of, for example, permanent facial nerve
palsy.
16.7.6 Limitations
Baseline and technical information regarding the procedure was missing for some
patients, limiting the ability to include them in multivariable modelling. Information about
the clinical effects and duration of CNP was limited in some patients. Multiple
comparisons in the univariable analysis raise the possibility of type 1 (false positive)
error, and no statistical correction was made for the number of tests carried out.
Patients were not randomized to particular techniques or medication regimens, and it is
likely that unmeasured confounders such as surgical expertise also influence the risk of
CNP or haematoma.
16.7.7 Conclusion
Long-term disability due to cranial nerve palsy following carotid endarterectomy
appears to be rare. As a result of this finding, CNP is unlikely to be included in
composite endpoints of future trials involving carotid surgery. However, CNP remains
common following CEA and is associated with haematoma. The results presented here
should give surgeons confidence to reassure patients that long-term disability due to
cranial nerve palsy following carotid endarterectomy is rare, to continue antiplatelet
therapy around the time of surgery, and to carefully consider the risk-benefit ratio for
revascularization in female patients.
Measurement of quality of life after CNP or haematoma might reveal that apparently
“minor” clinical outcomes have disproportionate effects on the patient’s well-being, and
future studies should focus on the quality of life impact of these complications.
116
17. Risk factors for stroke, myocardial infarction or death
following carotid stenting
Introduction 17.1
Carotid artery stenting was developed as a less-invasive alternative to carotid
endarterectomy for the prevention of TIA or stroke in patients with atheromatous
carotid stenosis. However, in interim analysis of the International Carotid Stenting
Study the risk of stroke, myocardial infarction or death within 30 days of
revascularization was higher in those patients who received their allocated CAS
procedure than in those who received their allocated CEA procedure (7.4% vs 4.0%,
p<0.01) [82].
Endovascular revascularization is a complex procedure requiring attention to pre-
procedure medication, haemodynamic control, stent design and optimal adjuvant
medical therapy [153]. Optimal stenting technique, patient selection and processes of
care for CAS have yet to be determined. For example, the use of cerebral protection
devices has become more common in recent years in the UK [97], but there is
conflicting evidence of their efficacy.
This analysis examines patients allocated to CAS in ICSS to determine whether there
are specific techniques, processes of care or baseline patient characteristics that are
associated with stroke, MI or death within 30 days of the procedure.
Methods 17.2
17.2.1 The stenting procedure
The trial protocol and inclusion/exclusion criteria for ICSS are summarised in Chapter
13 and presented in full in Appendix IV. Patients with more than 50% recently-
symptomatic carotid stenosis were eligible for randomization in the study if they were
felt suitable to undergo either CEA or CAS. Patients were excluded from enrolment in
the trial if the stenosis was due to non-atheromatous disease, if cardiac bypass was
planned within 1 month of the carotid procedure, or if there was previous stenting or
surgery on the symptomatic artery. Additionally, patients were specifically excluded
from CAS if their vascular anatomy was “unfavourable” – if there was vessel tortuosity
proximal or distal to the stenosis, visible thrombus, proximal CCA disease or “pseudo-
occlusion” (99% or more stenosis but with normal distal runoff on angiography).
117
Centres enrolling patients into ICSS were required to provide evidence that CAS
interventionists were sufficiently experienced in the procedure, having performed a
minimum of 50 stenting procedures and at least 10 of those in the carotid territory with
an acceptable complication rate. More inexperienced interventionists joined the trial in
probationary centres where procedures were supervised by a proctor until such time as
the required level of competency was achieved.
CAS in ICSS was performed in accordance with trial protocol: aspirin and clopidogrel
prior to the procedure was recommended, as was the use of a CPD whenever the
interventionist thought one could be safely deployed. CPDs and stents used at each
centre were approved by the ICSS steering committee and were CE-marked. The
administration of heparin during the procedure was mandatory. A CAS technical data
form was returned to the trial office by interventionists giving details of the equipment
and medications used in the procedure.
17.2.2 Outcome events
The combined endpoint of stroke, MI or death within 30 days of CAS is examined in
this analysis, and was reported to the trial office by investigators at each centre. For
each of these outcomes, confirmatory evidence was required (brain scan in the case of
stroke, ECG and / or cardiac enzyme levels for MI, death certificate and autopsy report
if available for death). Stroke was defined as “an acute disturbance of focal
neurological function with symptoms lasting more than 24 hours resulting from
intracranial vascular disturbance”. Two out of the three following criteria were required
for a diagnosis of MI – a clinical history of chest discomfort lasting more than 30
minutes, the development of specific ECG abnormalities consistent with MI or cardiac
enzyme levels more than twice the upper limit of the reference range. Two
investigators from the trial office adjudicated each event and then it was sent to a third,
independent, clinician for adjudication. Disagreements between investigators were
resolved by consensus after a second independent review.
17.2.3 Statistical analysis
A per-protocol analysis, including only those patients randomized to CAS in ICSS in
whom the stenting procedure was initiated, is presented in this chapter. Stenting was
deemed initiated if the patient received local or general anaesthesia prior to the
procedure. Patients crossing over to surgery or best medical therapy were therefore
excluded. Risk factors for stroke, MI or death within 30 days of the CAS procedure
118
were investigated sequentially in a binomial regression model. Patients with missing
data were excluded from each relevant analysis. The effect of cerebral protection type
and stent type could only be investigated in patients in whom these devices were
deployed, and the effect of post-stent dilation was examined only in those patients in
whom a stent was deployed. A multivariable model was developed using a forward
stepwise approach including variables potentially available for all CAS patients,
therefore excluding CPD type, stent type and post-dilation. A p-value of <0.05 was
considered “significant” in all analyses. Analyses were performed with Stata
(StataCorp. 2011. Stata Statistical Software: Release 12. College Station, TX:
StataCorp LP).
Results 17.3
17.3.1 Patient and procedure characteristics
1713 patients were randomized in ICSS, of whom 853 were allocated to CAS. In this
group, 828 procedures were initiated. Baseline characteristics of included patients are
presented in Figure 12 below. In summary, 70.4% of patients were male, and 53.5% of
patients were over the median age of 70 years. A stent was deployed in 752/816
(92.2%) procedures for which data were available. 367/816 (48.8%) of these
procedures used an open-cell design of stent, 371/816 (49.3%) used a closed-cell
stent. A CPD was deployed in 585/828 (70.6%) procedures, of which the majority
(n=464) were distal filter-type devices. 71.7% of patients were taking the combination
of aspirin and clopidogrel prior to the procedure.
17.3.2 Timing and cause of events
61/828 (7.4%) patients undergoing CAS suffered stroke, MI or death within 30 days of
the procedure. The events reported in these patients were 58 strokes, three myocardial
infarctions and one death unrelated to stroke or MI. One patient with stroke suffered a
subsequent (fatal) MI. 44/61 (72.1%) events occurred on the day of the procedure itself
(day 0), of which 21 were adjudicated to have occurred during the procedure. 21
occurred between the end of the procedure and day 1. In one patient a retinal infarct
was noted during follow-up and attributed to the procedure on day 0, but the timing in
relation to CAS was uncertain. In one further patient there was insufficient information
to determine the timing of the event.
Table 15 below summarizes the immediate cause or timing of the event, as reported by
the interventionist, in the 21 patients whose event was adjudicated to have occurred
119
during the procedure itself. The majority of these events (12/21) were precipitated by
the insertion, retrieval or use of endovascular equipment around the site of the lesion
itself.
120
Table 15. Reported immediate cause or timing of event in patients undergoing
CAS in ICSS
Precipitant Number of events (n=21)
Uncertain / undetermined 6
Post-dilation 5
Stent deployment 4
CPD manipulation 3
Intra-operative hypotension 2
Catheter angiography 1
121
17.3.3 Demographic and technical risk factors for stroke, MI or death within 30
days of the procedure
The results of univariable analysis of the influence of baseline demographic and
technical risk factors on the risk of stroke, MI or death within 30 days of CAS are
presented in Figure 12. The risk of the combined outcome increased 1.24 times for
each 5 years increase in age (95% CI 1.07 to 1.42, p<0.01). The risk was also
increased in patients with an open-cell carotid stent (RR 1.86, 95% CI 1.09 to 3.19,
p=0.02) and in patients with atrial fibrillation (RR 2.31, 95% CI 1.16 to 4.61, p=0.02).
The risk was lower in those who were current or former smokers at baseline as
compared to those who had never smoked (RR 0.25, 95% CI 0.10 to 0.63 and RR
0.77, 95% CI 0.46 to 1.29 respectively, global p<0.01) and those with amaurosis fugax
as the initial event prompting randomization in the trial as compared with those who
suffered stroke as the index event (RR 0.25, 95% CI 0.08 to 0.80 respectively, global
p<0.01). The risk was also lower in patients undergoing a right-sided procedure (RR
0.54, 95% CI 0.32 to 0.91, p=0.02). Patients in whom a CPD was used had a higher
risk of stroke, MI or death with nearly double the number of events, but this increase
was not statistically significant (RR 1.86, 95% CI 0.98 to 3.52, p=0.06).
Patients treated in centres who enrolled more than 50 patients in ICSS experienced a
lower risk of stroke, death or MI within 30 days of CAS (RR 0.60, 95% CI 0.37 to 0.96,
p=0.04), although the results in provisional centres where procedures were supervised
were not statistically different from those at more experienced centres.
122
Figure 12. Univariable predictors of the risk of stroke, myocardial infarction or death
in 828 patients undergoing CAS in whom the procedure was initiated
123
Figure 12 continued
124
One possible multivariable model investigating the effect of all variables potentially
available for all CAS patients (i.e. excluding those in whom a stent or CPD was not
deployed) is presented in Table 16. Independent predictors of the risk of stroke, MI or
death within 30 days of CAS were increasing age (RR 1.17 per 5 years increase, 95%
CI 1.01 to 1.37, p=0.04) and a right-sided procedure (RR 0.54, 95% CI 0.32 to 0.91,
p=0.02). The risk was significantly lower in patients taking the combination of aspirin
and clopidogrel prior to the procedure, as compared with any other antiplatelet regimen
(RR 0.59, 95% CI 0.36 to 0.98, p=0.04), and higher in those with retinal stroke as the
index event prompting randomization in the trial.
The model was reconstructed in a sensitivity analysis to exclude those patients in
whom a stent was not deployed. In this model, the use of an open-cell stent conferred
higher risk (RR 1.92, 95% CI 1.11 to 3.33, p=0.02).
125
Table 16. Independent predictors of risk of stroke, MI or death within 30 days of CAS in 748 ICSS per-protocol participants in whom the
procedure was initiated and for whom complete predictor data are available. Results obtained from multivariable binomial regression.
Variable Adjusted Risk Ratio (95% CI) P-value
Age (per 5 years increase) 1.17 (1.01 to 1.37) 0.04
Smoking status
Never smoked
Ex- smoker
Current smoker
1.00
0.86 (0.52 to 1.43)
0.33 (0.13 to 0.85)
N/A
0.55
0.02
Right-sided procedure 0.54 (0.32 to 0.91) 0.02
126
Table 16 continued
Variable Adjusted Risk Ratio (95% CI) P-value
Type of index event pre randomization in
ICSS
Stroke
Retinal Stroke
TIA
Amaurosis Fugax
1.00
2.47 (1.11 to 5.52)
0.99 (0.58 to 1.68)
0.32 (0.10 to 1.02)
N/A
0.03
0.96
0.06
Type of antiplatelet regimen
Any other antiplatelet
Aspirin and clopidogrel in combination
1
0.59 (0.36 to 0.98)
N/A
0.04
127
Discussion and conclusion 17.4
17.4.1 Summary
The risk of stroke, myocardial infarction or death within 30 days of carotid stenting in
ICSS was significantly higher in older patients, those undergoing a left-sided
procedure, in non-smokers, in patients who were not taking the recommended
combination of aspirin and clopidogrel and in those with retinal stroke as the index
event prompting randomization on the trial. Restricting the analysis to only those
patients in whom a stent was deployed revealed the use of an open-cell stent design to
be an independent predictor of increased risk compared with a closed-cell design.
17.4.2 Discussion
Evidence is building that CAS is less safe in older patients. Individual patient data
meta-analysis from the three European trials of CAS vs CEA found that although the
risks of stroke or death within 120 days of revascularization was similar between
patients allocated to CAS or CEA under the age of 70 years (5.8% vs 5.7%, RR 1.00,
95% CI 0.68 to 1.47), the risk of these complications with CAS were much higher in
patients 70 years or older (12.0% vs 5.9%, RR 2.04, 95% CI 1.48 to 2.82, interaction
p<0.01) [189]. There was no reduction in the severity of events in older patients. The
CREST study also showed a similar effect of age [84]. Age should therefore be taken
into account when selecting patients for the procedure, and the use of CAS in patients
over 70 years old should be largely reserved for those with a clear indication (e.g.
“hostile” surgical anatomy or post-radiotherapy).
Patients in ICSS in whom an open-cell stent was deployed were at higher risk of
complications, a finding echoed in other studies. A higher rate of ipsilateral stroke or
ipsilateral stroke death was found in patients treated with an open-cell stent in the
SPACE study (OR 2.13, 95% CI 1.07 to 3.76) [190], and similar trends have been
found by other groups [191] [192]. The effect of stent design may be more pronounced
in symptomatic patients because plaque is unstable, ulcerated or recently-complicated
with fresh thrombus which is stabilized by the greater coverage of closed-cell stents
with a smaller open area between cell struts, containing debris generated during
deployment of the stent to the site of the lesion. The structural difference between
open- and closed-cell stents is illustrated in Figure 13 below.
128
Figure 13. A scanning electron micrograph of closed-cell stent design before (a)
and after (b) expansion, and an open-cell stent before (c) and after (d)
expansion.
From [193] with kind permission from Lippincott Williams and Wilkins / Wolters Kluwer
Health
129
CPDs were developed for use in CAS in the assumption that they would reduce
embolism during stent deployment or lesion dilation. However, it is conceivable that the
passage of the device itself may produce particulate debris, and the deployment of a
filter could damage the distal vessel endothelium such that post-procedure thrombus
may form. The combination of findings in this chapter that many procedural minor
strokes were caused by manipulation of equipment around the site of the lesion itself,
and that certain types of patients were at higher risk of stroke, suggests that perhaps
both the generation of a shower of emboli during stenting and the brain’s ability to
tolerate these may be the clinical mechanism of causation in procedural stroke [194].
A systematic review comparing older case series with results after the introduction of
CPDs appeared to confirm the case for their use [195], but these results could be
improved by changes in patient selection and other developments in CAS technique
over time. The EVA-3S trial issued a clinical alert after randomizing 80 patients when
their finding that the 30-day risk of stroke in patients undergoing unprotected CAS was
four times that of those undergoing protected CAS [63], but the excess strokes in this
group may not have been down to CPD use alone, since only two strokes in those
treated without a CPD occurred on the day of the procedure itself. This analysis found
that CPD use in ICSS did not protect against stroke, MI or death within 30 days of the
procedure. Indeed ICSS patients enrolled in the ICSS-MRI Substudy in centres with a
general policy of CPD use experienced a 2.7-fold higher risk of new ischaemic brain
lesions seen on diffusion-weighted MRI following the procedure than those enrolled in
other centres where CPDs were not routinely used [64].
Two small randomized trials of CAS with and without filter-type CPDs have reported
results. One carried out in the United States showed no reduction in new DWI lesions
post-procedure in the CPD group [196], while the other which enrolled patients in the
UK demonstrated an increase in new ischaemic brain lesions post-procedure and
procedural micro-emboli as detected by transcranial Doppler in the CPD group [197].
The predominant design of CPD used in ICSS was a distal filter type, and newer
systems such as flow reversal devices may be more effective [65]. In the meantime,
the routine use of CPDs for the prevention of procedural stroke cannot be supported by
the results of ICSS and further research on newer devices is required to ensure their
safety and efficacy.
The reduction in 30-day neurological complications in patients taking combination
antiplatelet therapy to cover their CAS procedure has been described before [198], and
reinforces the recommendation that all CAS patients receive aspirin and clopidogrel
130
where appropriate to prevent thrombus forming on the metallic stent struts. The finding
that smoking was associated with a lower risk of complications is more difficult to
explain, and has not previously been described. It is therefore possibly a chance
finding, or one due to residual confounding.
It is interesting to contrast findings regarding the possible procedural precipitants of
stroke with findings from the Stenting and Aggressive Medical Management for
Preventing Recurrent Stroke in Intracranial Stenosis (SAMMPRIS) study [199]. This
trial involved potentially more hazardous intracranial artery stenting, but included those
with carotid and vertebral disease. Of 19 periprocedural ischaemic strokes, 12 were
thought to have been caused by perforator artery occlusion and 2 by delayed stent
occlusion. Embolism was only thought to be a contributing event in 5 patients. These
different mechanisms of stroke aetiology perhaps reflect the unique hazards of
intracranial artery stenting as compared with carotid stenting, but the contribution of
embolism to the mechanism of stroke in several patients highlights the extra danger of
stenting in symptomatic patients who may have fresh thrombus at the site of the
stenosis.
17.4.3 Limitations
The technology of carotid stenting is developing rapidly and findings related to the
specific stent types and device types used in ICSS may be superseded as newer
equipment enters common use. In some patients, information regarding baseline risk
factors was unavailable, and these patients were excluded from each relevant analysis.
Some procedures were abandoned before the deployment of a CPD or stent, and
these patients were not included in multivariable modelling which included as many
patients as possible with baseline predictor data. Multiple comparisons have been
made without statistical correction, raising the possibility of a Type I (false positive)
error. This is not a randomized comparison of stenting technique or processes or care,
and it is possible that unmeasured confounders are associated with the risk of stroke,
MI or death within 30 days of CAS. The limited number of outcome events in ICSS
limited the number of variables supported in a multivariable model.
17.4.4 Conclusion
Interventionists selecting patients suitable for CAS should take into account risk factors
for periprocedural stroke, MI or death that include the patient’s age, the side of the
procedure, and the nature of the event prompting intervention. Until such time as
131
consistently safe results can be demonstrated in older patients, stenting for those over
the age of 70 should have a clear indication (surgically-inaccessible stenosis, post-
radiotherapy or the patient’s wish to avoid surgery, for example). Those who require
revascularization of the left carotid artery should be aware of a potentially higher
procedural risk – although this of course should be balanced against the risk of stroke
without treatment. The results presented in this chapter favour the use of a closed-cell
stent without routine deployment of a cerebral protection device. They also reinforce
the need for dual antiplatelet cover with aspirin and clopidogrel in all patients.
132
18. Vascular anatomical factors influencing the risk of new
ischaemic brain lesions following carotid artery stenting
Introduction 18.1
Patients in the CAS arm of ICSS experienced a higher rate of stroke, MI or death within
30 days of the procedure than patients in the CEA arm, mostly attributable to an
excess of minor stroke [82]. Patients enrolled in the ICSS-MRI Substudy were
examined further with pre- and post-procedure brain scans to determine the number
and volume of new diffusion-weighted imaging (DWI) lesions – indicative of ischaemia
or infarction. 50% of patients allocated to CAS in the substudy were discovered to have
developed at least one new ischaemic brain lesion due to the procedure, compared
with just 17% of those allocated to CEA (OR 5.21, 95% CI 2.78 to 9.79, p<0.01) [64].
Lesions were more numerous in patients undergoing CAS, and were also more likely to
occur in cortical and adjacent white matter areas than those in CEA patients [200].
CAS is a technically demanding procedure that requires the passage of equipment
through the vascular tree up to and beyond the atherosclerotic lesion. There is
emerging evidence from the randomized clinical trials of CAS versus CEA that patients
undergoing CAS procedures performed by operators with low annual volumes of cases
may be at higher risk of stroke or death within 30 days of the procedure [126].
A number of vascular anatomical characteristics have been proposed by expert opinion
as contributing to a technically challenging procedure [123] [201]. This analysis
investigates the influence of baseline patient characteristics, procedural technique and
variation in vascular anatomy seen on pre-stenting catheter angiography on the risk of
new DWI brain lesions following CAS in patients enrolled into the ICSS-MRI Substudy.
Methods 18.2
18.2.1 The ICSS-MRI Substudy
The ICSS-MRI Substudy is discussed in Chapter 13 and the full protocol is supplied in
Appendix V. To be eligible for randomization in ICSS patients over 40 years old with
recently-symptomatic carotid stenosis were required to be suitable for revascularization
by either CAS or CEA. Importantly, they were excluded from randomization if they
presented with a stenosis that was known to be unsuitable for stenting due to pseudo-
occlusion (distal collapse of the artery with “string sign” indicating poor runoff from the
133
lesion), tortuous vascular anatomy, proximal common carotid artery stenosis or visible
thrombus on vascular imaging. CAS was carried out by accredited interventionists
according to trial protocol described in Appendix IV. CPD use was recommended
whenever the interventionist could “safely deploy” one, but was not mandatory.
Operators were required to use devices approved by the trial steering committee.
Patients who additionally enrolled in the ICSS-MRI Substudy underwent pre-procedure
brain MRI at either 1.5T or 3.0T field strength carried out up to a week before
revascularization. MRI was then repeated post-procedure at one to three days after
CAS and then at one month after CAS. Scan protocols contained diffusion-weighted
imaging (DWI) sequences to identify recent ischaemic brain lesions.
Patients underwent carotid stenting as per the ICSS protocol. This analysis includes
patients allocated to CAS in whom the stenting procedure was initiated and for whom a
baseline or pre-stenting catheter angiogram was available and of sufficient quality to be
analysed.
18.2.2 Analysis of brain and vascular imaging
A “lesion count” for each DWI sequence was obtained by the consensus reading of a
neurologist and neuroradiologist blind to the patient’s identity, the date and the nature
of the scan (e.g. pre- or post-procedural). The location (ipsilateral or contralateral) and
the total volume of lesions were also recorded.
Catheter angiograms showing vascular anatomy before stent insertion were evaluated
initially by one trained reader and then read again by the author to enable inter-rater
reliability to be calculated. Both readers were blind to the patient’s identity and the
selected outcome variable (number of new ischaemic brain lesions on post-procedure
DWI MRI). Standard definitions were agreed between the ICSS-MRI Substudy
investigators prior to the first scans being read. The degree of carotid artery stenosis
was calculated using the formula developed in NASCET [40], being the ratio of the
smallest diameter of the ICA at the level of the lesion divided by the distal ICA diameter
expressed as a percentage. ECA stenosis was measured in the same way, using the
distal ECA diameter. “Ulceration” and “irregularity” of the atherosclerotic plaque was
also defined according to the NASCET investigators’ methods [202], as illustrated in
Figure 14. “Pinhole” stenosis was defined as a luminal diameter too small to accurately
measure but with normal distal run-off in the vessel and no distal collapse. Many digital
copies of catheter angiography in the ICSS-MRI Substudy did not contain calibration
134
information, so “lesion length” was defined as the distance between proximal and distal
shoulders of the atherosclerotic lesion where the luminal diameter decreased to 20% of
its maximum expressed as a fraction of the common carotid artery on angiographic
views that most elongated that stenosis [203], as illustrated in Figure 15. A “diseased
CCA” was one that had the appearance of luminal stenosis or irregularity consistent
with atherosclerosis. The contralateral carotid artery was not always imaged in the
sequences available to the readers, so the measurement of contralateral carotid
stenosis provided by trial investigators at the time of the patient’s enrolment in ICSS
was used in the analysis. Images were prepared and analysed using Osirix MD,
(Pixmeo SARL, 2012).
135
Figure 14. Determination of smooth (D, E), irregular (C) or ulcerated (A, B)
atheromatous plaque at the carotid bifurcation
From [202] with kind permission from Lippincott Williams & Wilkins / Wolters Kluwer
Health
136
Figure 15. Measuring the length of maximal stenosis in three vascular anatomical profiles
CCA = common carotid artery, ICA = internal carotid artery, ECA = external carotid artery
Adapted from [203] under the terms of the Creative Common Attribution License. Original figure ©The Authors.
137
18.2.3 Statistical Methods
Each demographic, vascular or anatomical risk factor was investigated sequentially in a
negative binomial regression model, with the number of new DWI brain lesions on post-
procedure scan set as the dependent (outcome) variable. Comparisons between
groups are expressed as an odds ratio with 95% confidence interval, demonstrating
how many times more or less lesions occurred. Factors with statistical significance,
p<0.05, were considered for entry into a multivariable model constructed using a
forward stepwise approach. Patient with missing data were excluded from each
relevant analysis. Post-hoc sensitivity analyses for lesions ipsilateral to the treated
artery only as the outcome variable and for lesion volume as the outcome variable
were carried out in the same manner. Inter-rater reliability was calculated to produce a
kappa (κ) statistic and p-value, interpreted as proposed by Landis and Koch [204]. All
analyses were performed using SPSS version 22.0.0.0 (IBM Corp, 2013).
Results 18.3
18.3.1 Baseline patient characteristics
A total of 127/231 patients randomized in the ICSS-MRI Substudy were allocated to
CAS. Baseline (n=7) or pre-procedural (n=110) catheter angiograms with views
sufficient for analysis were available for 115 patients. The baseline demographic
characteristics and vascular risk factors of the patients included in this analysis are
shown in Table 17. Measured vascular anatomical features are shown in Table 18
below.
The median age of patients in this analysis was 71 years. 70% of patients were male. A
protection device (CPD) was used in 46/115 (40%) procedures, of which 91% were
distal filters and 9% were flow reversal systems. 53% procedures were on the right
carotid artery, 47% on the left.
138
Table 17. Demographic and procedural characteristics of 115 patients allocated to
CAS in the ICSS-MRI Substudy for whom adequate pre-procedure catheter
angiography was available
Parameter Number of patients (%) n=115
Right-sided procedure 61 (53%)
Cerebral protection device used 46 (40%)
Age (mean) 70.4 years
Female 34 (30%)
Treated hypertension 77 (67%)
Diabetes (any type) 22 (19%)
Treated hyperlipidaemia 71 (62%)
Current or former smoker 87 (76%)
Coronary heart disease 26 (23%)
Other cardioembolic source of thrombus 7 (6%)
Cardiac failure 3 (3%)
Peripheral arterial disease 21 (18%)
Contralateral carotid occlusion 7 (6%)
Contralateral carotid stenosis 70-99% 23 (20%)
139
Table 17 continued
Parameter Number of patients (%) n=115
Contralateral carotid stenosis 50-69% 12 (10%)
Contralateral carotid stenosis <50% (reference
group)
73 (64%)
Baseline mRS=4 2 (2%)
Baseline mRS=3 6 (5%)
Baseline mRS=2 28 (24%)
Baseline mRS=1 29 (25%)
Baseline mRS=0 (reference group) 50 (44%)
Cholesterol (mean) 4.9 mmol/l
Diastolic blood pressure (mean) 83.0 mmHg
Systolic blood pressure (mean) 157.3 mmHg
ARWMC (median) 4
Qualifying event was stroke 49 (43%)
Qualifying event was TIA 41 (36%)
Qualifying event was retinal TIA or retinal stroke 25 (22%)
140
Table 18. Vascular anatomical characteristics of 115 patients allocated to CAS in
the ICSS-MRI Substudy for whom adequate pre-procedure catheter
angiography was available
Parameter Value
Carotid artery NASCET degree of stenosis (median) 62.7%
Ratio of maximal stenosis length to CCA diameter (mean) 0.41 times
Angle between common and internal carotid arteries (median) 23.6 degrees
External carotid artery stenosis >50% 6 (5%)
Plaque ulceration 40 (35%)
Pinhole stenosis 2 (2%)
Common carotid artery atheroma 13 (11%)
141
18.3.2 Patient, procedural and vascular anatomical risk factors for new
ischaemic brain lesions on the post-procedural scan following CAS
62/115 (50.4%) of patients had at least one new DWI-MRI-positive lesion on their post-
procedure scan. The results of negative binomial regression analysis of the influence of
baseline demographic variables, vascular risk factors and vascular anatomical
characteristics on the number of new DWI-positive lesions on the post-procedure MRI
are summarized in Tables 19 and 20 below.
Factors that increased the risk of new DWI lesions on the post-procedure scan were
increasing age (OR 2.26 for each 10 years increase in age, 95% CI 1.74 to 2.89,
p<0.01), treated hypertension (OR 2.61, 95% CI 1.68 to 4.05, p<0.01), increasing
ARWMC (age-related white matter change score seen on baseline brain imaging) (OR
1.13 for each point increase, 95% CI 1.06 to 1.20, p<0.01) and stroke or TIA as the
qualifying event for enrolment into ICSS as compared with those who experienced a
retinal event (OR 4.30, 95% CI 2.46 to 7.53, p<0.01 for stroke and OR 1.91, 95% CI
1.06 to 3.43, p=0.03 for TIA).
Factors associated with a lower number of new DWI lesions on the post-procedure
scan were being female (OR 0.49, 95% CI 0.31 to 0.76, p<0.01), being a current or
former smoker (OR 0.48, 95% CI 0.31 to 0.76, p<0.01), cardiac failure (OR 0.17, 95%
CI 0.03 to 0.85, p=0.03) and 50-69% contralateral carotid artery stenosis as compared
with <50% contralateral carotid artery stenosis (OR 0.18, 95% CI 0.08 to 0.40, p<0.01).
Patients with a mRS score of 3 at baseline had a lower number of new lesions
compared with patients with no disability scoring 0 (OR 0.20, 95% CI 0.06 to 0.62,
p<0.01). The number of new DWI lesions was lower in patients undergoing right-sided
procedure (OR 0.57, 95% CI 0.38 to 0.85, p<0.01).
The vascular anatomical variables measured for this analysis – the degree of ipsilateral
carotid stenosis, the length of maximum stenosis, the CCA/ICA angle, plaque
ulceration, the presence of CCA atheroma or the presence of “pinhole” stenosis did not
significantly influence the number of new lesions.
142
Table 19. Univariable demographic and procedural predictors of new DWI-MRI
lesions following CAS in 115 patients in the ICSS-MRI Substudy for whom
adequate pre-procedure catheter angiography was available
Parameter p-
value
Odds Ratio
(OR)
95% confidence interval
for OR
Lower Upper
Right-sided stenosis <0.01 0.57 0.38 0.85
Cerebral protection device used 0.22 1.30 0.83 1.88
Age (per 10-year increase) <0.01 2.26 1.74 2.89
Female <0.01 0.49 0.31 0.76
Treated hypertension <0.01 2.61 1.68 4.05
Diabetes (any type) 0.07 1.59 0.97 2.61
Treated hyperlipidaemia 0.69 0.92 0.61 1.38
Current or former smoker <0.01 0.48 0.31 0.76
Coronary heart disease 0.43 1.21 0.76 1.94
Other cardioembolic source of
thrombus
0.28 1.56 0.69 3.49
Cardiac failure 0.03 0.17 0.03 0.85
Peripheral arterial disease 0.29 0.75 0.45 1.27
Contralateral carotid occlusion 0.21 1.69 0.75 3.80
143
Table 19 continued
Parameter p-
value
Odds Ratio
(OR)
95% confidence
interval for OR
Lower Upper
Contralateral carotid stenosis 70-
99%
0.12 0.66 0.39 1.11
Contralateral carotid stenosis 50-
69%
<0.01 0.18 0.08 0.40
Contralateral carotid stenosis <50%
(reference group)
N/A 1 N/A N/A
Baseline mRS=4 0.14 0.26 0.04 1.57
Baseline mRS=3 <0.01 0.20 0.06 0.62
Baseline mRS=2 0.20 1.38 0.84 2.26
Baseline mRS=1 0.40 0.81 0.49 1.33
Baseline mRS=0 (reference group) N/A 1 N/A N/A
Cholesterol (1 mmol increase in
serum total cholesterol)
0.30 0.89 0.71 1.11
Diastolic blood pressure (per
10mmHg increase)
0.13 0.90 0.98 1.00
Systolic blood pressure (per
10mmHg increase)
0.98 1.00 0.91 1.09
144
Table 19 continued
Parameter p-value Odds Ratio
(OR)
95% confidence interval for OR
Lower Upper
ARWMC (per unit
increase)
<0.01 1.13 1.06 1.20
Qualifying event was
stroke
<0.01 4.30 2.46 7.53
Qualifying event was TIA 0.03 1.91 1.06 3.43
Qualifying event retinal TIA
or retinal stroke (ref.
group)
N/A 1 N/A N/A
145
Table 20. Univariable vascular anatomical predictors of new DWI-MRI lesions
following CAS in 115 patients in the ICSS-MRI Substudy for whom adequate
pre-procedure catheter angiography was available
Parameter p-
value
Odds
Ratio (OR)
95% confidence
interval for OR
Lower Upper
Carotid artery NASCET degree of
stenosis (per 10% increase)
0.69 0.97 0.83 1.13
Ratio of maximal stenosis length to
CCA diameter (per unit increase)
0.46 1.28 0.67 2.45
Angle between common and internal
carotid arteries
0.12 0.99 0.98 1.00
External carotid artery stenosis >50% 0.08 0.42 0.16 1.10
Plaque ulceration 0.66 1.10 0.73 1.66
Pinhole stenosis 0.88 1.13 0.25 5.05
Common carotid artery atheroma 0.69 0.88 0.47 1.65
146
Factors in the univariable analysis with significance <0.05 were considered for entry
into a multivariable model. One possible model is illustrated below in Table 21. Factors
independently increasing the number of new DWI brain lesions on the post-procedure
scan in this model are a left-sided procedure (OR 1.59, 95% CI 1.04 to 2.44, p=0.03),
increasing age (OR 2.10 per 10 years increase, 95% CI 1.61 to 2.74, p<0.01), being
male (OR 2.83, 95% CI 1.72 to 4.67, p<0.01), treated hypertension (OR 2.04, 95% CI
1.25 to 3.33, p<0.01) and absence of a diagnosis of cardiac failure (OR 6.58, 95% CI
1.23 to 35.07, p=0.03).
Inter-rater agreement, calculated for binary anatomical characteristics, was moderate
for CCA atheroma (κ=0.50, p<0.01), substantial for the type of stenosis (smooth,
irregular or ulcerated, κ=0.76, p<0.01) and slight for pinhole stenosis (κ=0.13, p=0.06).
147
Table 21. Independent predictors of new DWI-MRI lesions following CAS in 115
patients in the ICSS-MRI Substudy for whom adequate pre-procedure catheter
angiography was available
Parameter p-
value
Odds
ratio
(OR)
95% confidence interval for OR
Lower Upper
Left-sided stenosis 0.03 1.59 1.04 2.44
Age (per 10-year
increase)
<0.01 2.10 1.61 2.74
Male <0.01 2.83 1.72 4.67
Treated hypertension <0.01 2.04 1.25 3.33
Absence of cardiac
failure
0.03 6.58 1.23 35.07
148
Discussion and conclusions 18.4
18.4.1 Summary
Of the vascular anatomical characteristics measured specifically for this analysis, none
significantly influence the risk of new DWI-positive lesions on the post-procedure scan.
Independent factors increasing the number of new lesions were age, approximately
doubling risk for each advancing decade of life, and treated hypertension. A diagnosis
of cardiac failure, being female and undergoing a right-sided procedure all
independently decreased the number of new lesions.
18.4.2 Discussion
The assumption that a difficult procedure – technically challenging for an inexperienced
investigator – will be associated with an increased procedural risk has led to attempts
to define which vascular anatomical variants require extra skill to navigate during CAS.
A Delphi panel of CAS experts proposed a scoring system for rating difficulty based on
vessel angulation, tortuosity of the arteries, the presence of external or common carotid
artery disease or aortic arch variants [123]. Studies focussing on clinical endpoints,
usually stroke or death, have found a number of vascular anatomical factors that
increase the risk of an adverse outcome – lesion length [203] [122] [120], plaque
ulceration or plaque calcification [122], aortic arch variants [122] [205], angulation of
the ICA-CCA bifurcation [120] and the side of the procedure [120]. Atypical aortic arch
configuration may also be linked to longer catheter manipulation times, which itself may
be linked to adverse outcome [206]. However, other authors have failed to find a link
between lesion or vascular characteristics and outcome [207].
Factors specifically associated with an increased risk of new DWI-MRI lesions following
CAS in other studies include age [208], the presence of an ulcerated stenosis [208],
aortic atherosclerotic lesions [209] and increasing lesion length [208].
New ischaemic brain lesions following CAS are of concern. Subclinical injury to the
brain from microemboli generated during the procedure, i.e. insufficient to cause
clinically-apparent focal neurological symptoms, is thought to be a risk factor for
cognitive dysfunction perhaps too subtle to be detected without neuropsychological
testing [210]. CAS was associated with a larger decrease in cognitive ability than CEA
in selected patients enrolled at two ICSS centres, although the difference was not
statistically significant and the sample size was relatively small compared to the overall
149
trial size [211]. A decrease in cognitive performance in both groups occurred
regardless of the baseline white matter lesion burden of the patients studied [212].
This analysis demonstrated a higher number of new DWI-MRI lesions following CAS in
patients undergoing left-sided procedures, but was not able to replicate the findings of
others of an increased procedural risk with pinhole stenosis, length of stenosis or ICA-
CCA angulation. In ICSS the anatomical suitability of the stenosis for CAS or CEA (the
patient was required to be suitable for either to be included in the trial) was determined
prior to randomization using non-invasive imaging including duplex ultrasound or CT or
MR angiography. Pre-stenting catheter angiograms were primarily obtained to plan
procedures, and thus the films or DICOM images available for this analysis did not
routinely include the contralateral carotid artery or the aortic arch and CCA origin. It is
possible that patients thought to be “high risk” for stenting because of their vascular
anatomy were excluded from enrolment in ICSS. The trial protocol specified tortuous
anatomy proximal or distal to the stenosis, the presence of visible thrombus, proximal
CCA stenotic disease or pseudo-occlusion as examples of high risk anatomies. The
relatively low percentage of patients with pinhole stenosis and the low average CCA-
ICA angle in our study may reflect this. Inter-rater agreement for some of the measured
features, such as plaque ulceration, has also been shown to be poor [202].
Given the lack of inter-rater agreement for binary features, and the lack of a strong
association between vascular characteristics and the outcome, this data does not
support selecting patients for CAS based on “safe” vascular anatomy. In future, newer
imaging techniques such as 3T MRI looking at the composition of carotid plaque may
prove more useful in risk stratification if they are able to accurately identify fragile
plaques vulnerable to fragmentation and dislodgement by a vascular catheter [213].
Older patients in the ICSS-MRI Substudy were found to have a higher number of new
ischaemic brain lesions after stenting, and the trial also demonstrated as association
between the total volume of DWI lesions and hemispheric stroke in this patient group
[64] [200]. This finding agrees with clinical data from the Carotid Stenting Trialists’
Collaboration (CTSC) which conducted a meta-analysis of individual patient data from
ICSS [82], EVA-3S [79] and SPACE [77]. This revealed a doubling of risk of stroke,
myocardial infarction or death within 30 days of CAS compared with CEA in patients
over the age of 70 years with no significant difference in the risk in younger patients.
The mechanism by which age increases risk is not apparent from this analysis, but
could be related to an increase in the overall burden of atheroma in the vasculature
over time, a greater baseline ARWMC score and therefore decreased “cerebral
150
reserve” when ischaemia occurs, or age may be a marker of decreased physiological
reserve to tolerate the procedure and hospitalization. An increase in atheroma in the
aorta with age may also go some way to explaining concerns about the occurrence of
new DWI lesions on MRI brain following catheterization of the coronary arteries where
the carotid artery is not manipulated [214].
That cerebral protection devices did not protect against the occurrence of new DWI
brain lesions in our analysis agrees with the findings of two small randomized studies
of protected vs unprotected CAS [196] [197], but disagrees with one other study and
one systematic review which found a reduction in embolic load with CPD use [215]
[192]. However, even these two analyses demonstrate that new DWI lesions still
frequently occur following protected CAS. Most CPDs in ICSS and the ICSS-MRI
Substudy were of the distal filter type, and newer flow-reversal devices may offer better
protection against distal embolism [66].
18.4.3 Limitations
Views on catheter angiography were often limited, and thus it was not possible to
evaluate the anatomy of the intracranial portion of the ICA, the proximal CCA or the
anatomy of the aortic arch. All of these may be important areas of technical difficulty,
and could be further studied on CT or MR angiogram in the ICSS-MRI Substudy or
other patient cohorts. There is no correction in this analysis for multiple statistical
comparisons, and there is therefore a possibility of type I (false positive) error in
interpreting a positive result.
18.4.4 Conclusions
It is unlikely that the excess of new DWI brain lesions seen in CAS patients in the
ICSS-MRI Substudy was due to vascular anatomical variation alone. Patients who
match the inclusion criteria for ICSS and in whom stenting is thought viable need not
be excluded from the procedure on the basis of individual vascular anatomical
characteristics as assessed on catheter angiogram. In agreement with other studies,
age remains a risk factor for CAS and should be taken into account when deciding
between CAS and CEA. Caution should be exercised in offering CAS to patients over
the age of 70, as discussed in Chapter 17. There is mounting evidence that filter-type
CPDs do not offer protection against brain ischaemia during CAS, and our findings
therefore do not support their routine use.
151
19. Validation of an existing risk score for carotid stenting
Introduction 19.1
Patients allocated to stenting in the International Carotid Stenting Study experienced a
higher risk of stroke than patients allocated to endarterectomy in an analysis limited to
those whose procedures were initiated (7.0% vs 3.3%, p<0.01). This was mostly
attributable to an excess of minor (non-disabling) stroke [82]. Matching this finding,
three times the number of patients allocated to CAS in the ICSS-MRI Substudy were
found to have new ischaemic brain lesions on post-treatment MRI scans compared
with those undergoing endarterectomy [64]. CAS patients not only had higher numbers
of ischaemic lesions after revascularization, but these were more likely to occur in
cortical tissue and sub-adjacent white matter areas [200]. Patients who developed
clinical signs of hemispheric stroke had a higher overall lesion volume than those who
did not [64]. “Silent” brain infarction has been associated with dementia and cognitive
decline in other patient cohorts [216], and new lesions are thus of concern in the
context of carotid revascularization.
A risk score for the development of new ischaemic brain lesions following CAS,
constructed by Gröschel et al, was developed from a dataset of 176 patients
undergoing CAS at one centre from 2000-2006 who underwent diffusion-weighted
brain MRI sequences before and up to 48 hours after the procedure. Just over 51% of
their patients were symptomatic, and all had ≥70% carotid stenosis. Age over 70 years,
the presence of an ulcerated stenosis and lesion length over 1cm were identified as
risk factors for any new ipsilateral lesion on diffusion-weighted MRI following CAS
[208]. An analysis of the performance of this risk score is presented in this chapter.
Other risk prediction scores for CAS complications have been developed, but are not
studied here as they focus on clinical endpoints [122], address the choice between
CAS and CEA [217] [171], use MR plaque imaging techniques not in routine use in
ICSS [217] or focus on “high risk” patients not suitable for surgery [218] that would not
meet the inclusion criteria for the ICSS-MRI Substudy.
152
Methods 19.2
19.2.1 The ICSS-MRI Substudy and anatomical feature rating
The ICSS-MRI Substudy protocol is summarized in Chapter 13 and described in
greater detail in Appendix V. In summary, patients over the age of 40 with greater than
50% symptomatic carotid stenosis and capable of giving informed consent to trial
participation were randomized in ICSS to either CAS or CEA. CAS was carried out
according to ICSS trial protocol [81], including the use of a cerebral protection device
when the operator thought it safe to deploy one. It was recommended that patients be
prescribed aspirin and clopidogrel to cover the procedure.
Patients who additionally enrolled in the ICSS-MRI Substudy underwent pre-procedural
MRI at 1.5T or 3.0T field strength up to 7 days before CAS with repeat sequences one
to three days after the procedure. Scan protocols included diffusion-weighted
sequences to demonstrate areas of acute of subacute brain ischaemia or infarction.
The presence or absence of at least one new ipsilateral DWI-positive lesion was
decided by the consensus reading of a neurologist and neuroradiologist. Vascular
anatomical features – lesion length and the presence or absence of ulceration – were
recorded by a trained reader in a separate analysis blind to the selected outcome of
any new ipsilateral ischaemic brain lesion on the post-procedure scan as detailed in
Chapter 18.
Patients in the ICSS-MRI Substudy allocated to CAS in whom the procedure was
initiated, and for whom sufficient baseline or procedural catheter angiography was
available, were studied in this analysis. A score was allocated to each patient in the
analysis according to the characteristics in Table 22 below. The patient’s age in years
on the day of enrolment in ICSS is used in this analysis. Many digital angiograms did
not contain calibration information to directly measure the length of stenosis. Therefore
the “total stenosis length” was defined as the distance between the most proximal and
distal shoulders of the lesion, expressed as a fraction of the common carotid artery
diameter as illustrated in Figure 16. This was then multiplied by an assumed “average”
common carotid diameter of 0.65cm [142] to give a total length of stenosis in
centimetres. “Ulceration” was defined, as proposed by the NASCET investigators, as “a
crater penetrating into a stenotic plaque and double density on ‘en face’ view” [202],
illustrated in Figure 14 (Chapter 18). DICOM images were evaluated using Osirix MD,
(Pixmeo SARL 2012).
153
Figure 16. Measuring the total length of stenosis in three vascular anatomical profiles
CCA = common carotid artery, ICA = internal carotid artery, ECA = external carotid artery
154
Table 22. Allocation of risk points in the Gröschel (2008) risk score
Study Risk factors Risk points
Gröschel et al.
(2008) [208]
Age ≥ 70 years 1
Age ≥ 80 years 1
Lesion length > 1cm 1
Ulcerated stenosis 1
155
19.2.2 Statistical Methods
The proportion of patients with at least one new ipsilateral DWI-positive lesion on post-
procedure imaging was calculated for patients in each score bracket, expressed as a
percentage. A 95% confidence interval is given for each proportion. Differences
between the proportions of patients with a new DWI lesion in each group was assessed
by means of a chi-squared (χ2) test for trend. A receiver operator curve was
constructed to assess the performance of the score in our dataset, and the area under
the curve calculated with 95% confidence interval. In all analyses p<0.05 was assumed
to confer statistical significance. Analyses were performed using GraphPad (GraphPad
Software, 2013. GraphPad Instat version 3.00 for Windows 95. San Diego, California:
www.graphpad.com).
Results 19.3
Baseline characteristics of patients included in this analysis are presented in Tables 17
and 18 (Chapter 18). In summary, the mean age was 70.4 years, and 70.4% of patients
were male. 52/115 (45.2%) patients had at least one new DWI-positive lesion ipsilateral
to the treated artery of post-procedure MRI.
Using an assumed diameter of 0.65cm, the mean length of stenosis was 1.35cm.
18/115 (15.7%) patients had an ulcerated lesion.
33.3% of patients with a Gröschel score of 0 had at least one new ipsilateral ischaemic
brain lesion on the post-procedure MRI scan, 36.6% of those with a score of 1, 51.6%
of those with a score of 2, 61.1% of those with a score of 3 and 75.% of those with a
score of 4. The numbers of patients in each group and 95% confidence interval for the
proportion with a new ischaemic lesion is presented in Table 23. The observed
increase in patients with an increasing Gröschel score was statistically significant
(p=0.02, χ2 test for trend).
The receiver operator curve for the performance of the Gröschel risk score is presented
in Figure 17 below. The area under the curve is 0.62 (95% CI 0.52 to 0.73, p=0.02).
156
Table 23. Results of application of the Gröschel (2008) risk score to 115 patients undergoing CAS in the ICSS-MRI Substudy
Gröschel
Score
Number of patients with
no new ipsilateral DWI-
positive lesions following
after CAS (n)
Number of patients with new
ipsilateral DWI-positive
lesions following CAS (n)
Proportion of patients with
new ipsilateral DWI-positive
lesions following CAS (%)
95% CI for proportion of patients
with new ipsilateral DWI-positive
lesions following CAS
0 14 7 33.3% 17.1% to 51.4%
1 26 15 36.6% 23.6% to 51.9%
2 15 16 51.6% 34.8% to 68.0%
3 7 11 61.1% 38.5% to 79.8%
4 1 3 75.0% 28.9% to 96.6%
χ2 test for trend p=0.02
157
Figure 17. Receiver-operator curve for the application of the Gröschel (2008) risk
score in 115 patients undergoing CAS in the ICSS-MRI Substudy
Darker line represents receiver operator curve
Straight line represents area under curve =0.5
158
Discussion and conclusions 19.4
19.4.1 Summary
45.2% of patients in the ICSS-MRI Substudy who were allocated to CAS, and in whom
adequate baseline or pre-procedure catheter angiography was available, suffered at
least one new ipsilateral ischaemic brain lesion on the post-procedure DWI-MRI scan.
While the Gröschel score distinguished between lower-risk and higher-risk groups (χ2
test for trend p=0.02), the sensitivity and specificity of the score in a receiver operator
curve was limited (AUC 0.62, 95% CI 0.52 to 0.73, p=0.02).
19.4.2 Limitations
The length of stenosis as measured in our study is an approximation – calibration
information was not contained in many digital angiograms, so to maintain consistency
we use the diameter of the CCA to derive a calculated length. However, the diameter of
the CCA has previously been found to vary by the side of the artery, sex, age, body
surface area, blood pressure and neck length [142] and in fact may vary as much as
14% within individuals depending on the phase of the cardiac cycle during which it is
imaged [219]. Patients randomized in the ICSS-MRI Substudy were required to be
equally suitable for both CAS and CEA, and it is possible that patients with very long or
ulcerated stenosis were excluded from randomization if interventionists thought them
unsuitable for CAS because of a high anticipated procedural risk.
19.4.3 Discussion
The area under the receiver operator curve (AUC) in this analysis (0.62) compares
unfavourably with the original performance of this score in the patient population from
which it was derived, which was 0.71 (0.76 for patients treated after the year 2004 only,
n=75) [208]. It should be noted that as the AUC corresponds to sensitivity and
specificity cut-offs for discriminating between those with or without new ischaemic brain
lesions, it does not necessarily indicate the level of clinical relevance [220]. The high
proportion of patients suffering new lesions in the lowest risk group in the ICSS-MRI
Substudy suggests that the Gröschel score is not sufficiently sensitive to be clinically
useful in classifying patients at risk of MRI-detected ischaemic brain lesions after
following CAS.
Chapter 18 illustrates that while age was a significant risk factor for the occurrence of
new ischaemic lesions in the ICSS-MRI Substudy, ulceration and the length of the
159
target lesion appeared to have little influence. This finding suggests that the
performance of the score in this patient cohort was mainly driven by the effect of age
on the outcome. Therefore it is more likely that the key factor in the formation of new
ischaemic brain lesions during the procedure is the ability of the brain to tolerate
microemboli, rather than an increased risk of generation of debris during the procedure.
Indeed, the clinical outcome of CAS also appears to be strongly influenced by the age
of the patient, with a higher rate of stroke in this patient population [221] [222] [189]
[116] [223]. Reduced physiological reserve to tolerate the procedure, medications and
hospitalization, a higher burden of white matter disease with decreased cerebral
reserve to tolerate emboli, or an association between ageing and anatomical features
increasing the complexity of intervention [224] are all possible explanations for this
effect.
As detailed in Chapter 18, inter-rater agreement for the detection of carotid plaque
ulceration was limited in the ICSS-MRI Substudy, and other authors have reported little
agreement between angiographic appearances and subsequent histology [202] [225].
However, it should also be noted that while lesion ulceration was not a strong predictor
of subsequent brain ischaemia in the ICSS-MRI Substudy the risk of stroke without any
revascularization procedure may be higher than in patients with smooth lesions.
Patients enrolled in NASCET with a high degree of stenosis and an ulcerated lesion
experienced an up to 73% risk of subsequent ipsilateral stroke within 24 months
compared to a 21% risk in those patients without ulceration of the plaque [226].
Likewise, increasing age is a risk factor for stroke on medical therapy alone for
symptomatic carotid stenosis [43]. Thus the Gröschel score tested in this chapter may
give an indication of the risk of intervention, but it does not give an indication of the
likely risks of stroke on medical treatment alone or the likely risk of CEA, and therefore
is of limited use in selecting an overall treatment strategy in patients with greater than
50% symptomatic carotid stenosis.
19.4.4 Conclusions
Although there was an increase in the number of new ischaemic brain lesion on post-
procedure MRI after CAS with increasing Gröschel score, this risk score did not exhibit
adequate discrimination between low- and high-risk groups for routine clinical use, in
part because 33.3% of the lowest-risk group still suffered at least one new ipsilateral
ischaemic brain lesion on the post-procedure scan. Its routine use for selection of
patients for CAS cannot be recommended without validation in further patient groups.
160
Age is a strong predictor of both clinical outcome and “silent” brain embolism, and
should continue to be taken into account when selecting patients for carotid stenting.
161
20. Discussion, conclusions and future directions
Discussion of key findings 20.1
20.1.1 Carotid endarterectomy
This thesis presents several key findings regarding the safety of carotid
endarterectomy. Most notably, women had around double the risk of the combined
endpoint of stroke, myocardial infarction or death and nearly double the risk of the
combined risk of any major complication (stroke, MI, CNP, haematoma or death)
following CEA compared with men. That women might experience a greater operative
risk has been described before [98], and this is of particular concern because of their
lower risk of stroke on medical therapy alone – hence the observation in ECST that the
number of patients needed to undergo surgery to prevent one ipsilateral stroke over 5
years was just nine in men against 36 for women [105]. Surgeons should carefully
consider the risk-benefit calculation for revascularization in this patient group.
Although women experienced a higher risk of cranial nerve palsy in ICSS, the risk of
persisting disabling CNP was around 1 in 1000, with most CNPs resolving within 30
days. The association between haematoma and CNP suggests that careful attention to
haemostasis may reduce the latter’s incidence. Although patients should be warned
about possible symptoms of CNP, they can also be reassured of a generally good
eventual outcome even when CNP does occur, and this complication should not
dissuade surgeons from offering revascularization.
There is evidence that in the last few years the short-term risks associated with CEA
may have fallen slightly [227]. The risk of surgery was acceptably low in ICSS – only
3.3% of patients in whom the allocated CEA procedure was initiated suffered a stroke
within 30 days of the procedure – and the only independent predictor of the risk of
complications was baseline diastolic blood pressure. Attention to blood pressure in
symptomatic patients undergoing revascularization might reduce subsequent
complications, but this should be managed in the clinical context with particular care in
patients who have contralateral carotid stenosis or vertebral artery stenosis.
It is not surprising that risk scores derived from other patient populations did not
adequately discriminate between low- and high-risk groups of ICSS patients given the
lack of strong predictors of risk. Nonetheless, most published surgical risk scores have
162
not previously been validated, so to be of use in clinical decision-making they should
be tested in other patient cohorts or updated.
20.1.2 Carotid stenting
This thesis identifies advancing age as a risk factor for complications following CAS,
but not for CEA, in agreement with individual patient meta-analysis across the
European trials of CAS vs CEA which suggested that CAS was as safe as CEA in
younger patients [115].
The finding that cerebral protection devices – mostly distal filters – offered no
statistically significant protection from stroke, myocardial infarction or death or the
occurrence of new ischaemic brain lesions following CAS in ICSS is disappointing, but
concurs with the results of other randomized comparisons [196] [197]. At best the
benefit of their deployment in CAS is unproven, and cannot be recommended as
routine. Similarly, results were disappointing in those patients treated using an open-
cell stent and the findings in this thesis favour the use of a closed-cell stent.
Inexperienced CAS interventionists are sometimes advised to avoid undertaking
stenting in patients with “challenging” vascular anatomy [123] thought to make the
procedure more technically difficult. In Chapter 18, this thesis argues that the influence
of patient characteristics, such as age, male sex and hypertension, exert a greater
influence on the number of new post-procedure ischaemic lesions as measured on MRI
brain, although patients with a left-sided stenosis experience approximately double the
number of new lesions. While an important observation, it must be kept in mind that
patients with lesions thought “unsuitable” for stenting were excluded from
randomization in ICSS, and that more inexperienced operators who performed trial
procedures were supervised. Indeed, patients treated in supervised centres in ICSS
experienced a lower risk of stroke, MI or death following CAS [82].
Conclusions 20.2
The case for offering carotid revascularization in suitable patients with ≥50%
symptomatic carotid stenosis due to atherosclerosis, who would otherwise experience
a high risk of recurrent stroke on medical therapy alone, is clear. Some groups of
patients, however, at are increased risk from intervention. In isolation, the subgroup
analyses presented in this thesis should be interpreted with caution. Yet combined with
data from individual patient meta-analysis of the European trials of CAS vs CEA, and
risk modelling data from the original NASCET and ECST trials of carotid surgery versus
163
medical therapy, there is evidence from the work presented in this thesis that in the
short term stenting is more hazardous in the elderly, that the routine use of filter-type
cerebral protection devices in CAS is not supported by high-quality evidence, that
closed-cell stents are safer than open and that women should be informed of a
potentially small increased risk of stroke, MI, cranial nerve palsy, haematoma or death
following endarterectomy.
What is not addressed in these analyses is whether the overall benefits of intervention
extend to lower-risk patients, and it is this question that is the subject of ongoing
research.
Future directions – who should have what treatment? 20.3
20.3.1 Revisiting the case for intervention
The standard of care for clinically stable patients with greater than 50% symptomatic
carotid stenosis has been to perform carotid endarterectomy (in preference to carotid
stenting in low-risk patients) to remove the symptomatic plaque and thus reduce the
long-term risk of recurrent stroke or TIA [128] [153] [130]. However, since the landmark
large randomized trials of CEA vs medical therapy, NASCET and ECST, there have
been a number of developments in care that call into question the benefit of
intervention in all patients with this degree of stenosis. This thesis has explored the
subject of which patients are at risk of complications from CEA or CAS, but in order to
select patients who will benefit in the long-term from revascularization it is necessary to
consider their risk of stroke without intervention as well.
Medical therapy for secondary prevention of stroke has changed. Previous trials of
CAS or CEA for carotid stenosis did not specify targets for control of vascular risk
factors, but numerous guidelines now exist for primary and secondary prevention
therapy [228] [229] [230] [231]. Cholesterol-lowering statins were not available during
recruitment into NASCET, but their use in secondary prevention for stroke is now
widespread in the knowledge that they reduce the risk of stroke in patients with carotid
disease [232] [233].
The pathology of symptomatic carotid plaque is well-described. An inflamed plaque
with macrophage infiltration, fresh thrombus, a large lipid core and rupture of the
fibrous plaque are all histological features associated with stroke (as opposed to TIA or
amaurosis fugax) [234]. Catheter angiography, the method used to determine the
degree of stenosis for entry to the NASCET and ECST trials, has fallen out of common
164
use as newer less invasive imaging techniques have been developed – ultrasound, CT
and MR angiography and advanced plaque imaging with MRI or 3D ultrasound. The
degree of stenosis on its own is currently used in the selection of patients to undergo
revascularization and is one potential marker of plaque instability – larger plaques
causing higher degrees of stenosis are more likely to cause symptoms due to a more
unstable plaque composition with less fibrous tissue. The ability to directly image
plaque components associated with high stroke risk such as a lipid-rich core [235]
[236], ulceration, thin or ruptured fibrous cap [235] [236] or intra-plaque haemorrhage
[235] [237] [236] with MRI is beginning to change the way patients are risk stratified
and may blur the line between asymptomatic and symptomatic patients when selecting
patients for revascularization based on their risk of stroke. Work is underway, for
example, in the Plaque At Risk (PARISK) study to enrol 300 patients with <70%
recently-symptomatic carotid stenosis, who are not scheduled to undergo carotid
endarterectomy. These patients will undergo detailed MR, CT and US of the carotid
arteries to determine which plaque characteristics confer a high risk of recurrent stroke
[238].
20.3.2 The risks of intervention
Detailed plaque imaging may similarly give clues as to the risk of the revascularization
procedure itself. Echolucent plaques on ultrasound – unstable, lipid-rich or
haemorrhagic plaques more likely to cause neurological symptoms untreated have
been shown to increase the risk of carotid stenting, perhaps through the dislodgement
of adherent thrombus during the passage of endovascular equipment across the lesion
[239]. Ongoing randomized trials with a carotid intervention arm such as ECST-2,
described below, are incorporating detailed modern carotid imaging into their baseline
assessment of patients and will use detailed post-procedure MRI to examine the
occurrence of procedural brain ischaemia, opening up the possibility of further
characterising the risk of intervention based on radiological plaque appearance.
Collaborative work continues to establish the determinants of the risks of intervention,
pooling individual patient data from the European trials of carotid stenting and surgery
through the Carotid Stenting Trialists’ Collaboration [240]. This larger dataset, which
includes data from patients enrolled in ICSS, has examined the effect of baseline
patient characteristics on procedural risk as described in Chapters 14 and 17, and has
recently published an analysis of the effect of operator experience on the outcome of
carotid stenting, demonstrating a more favourable outcome in those interventionists
that performed more than five procedures per year within the trials [126].
165
20.3.3 Ongoing clinical trials
The second European Carotid Surgery Trial (ECST-2) [172], designed by many of the
trialists involved in ICSS, intends to answer some of these questions about how best to
treat patients with carotid stenosis. It is a prospective randomized trial of patients with
low-to-intermediate risk carotid stenosis ≥50%, enrolling those with asymptomatic
plaques and recently-symptomatic patients whose calculated risk of ipsilateral stroke is
less than 15% over 5 years. The Carotid Artery Risk Score used to calculate risk is
adapted from an algorithm designed at Oxford University taking into account age, sex,
the degree of stenosis, the nature of any symptomatic event and how recently it
occurred, vascular risk factors and plaque morphology [241]. Patients in a medical
therapy group will receive “optimized medical therapy” (OMT) – antiplatelet therapy,
target-led control of cholesterol level and blood pressure – while an intervention group
will receive OMT plus revascularization. The trial tests the hypothesis that patients with
a low predicted stroke risk on OMT alone will not benefit from revascularization due to
periprocedural complications.
SPACE-2 (Stent-Protected Angioplasty in Asymptomatic Carotid Artery Stenosis
versus Endarterectomy Trial-2) [173] was originally designed as a randomized trial for
patients with asymptomatic carotid stenosis ≥70% as measured by ultrasound. Its
protocol called for a recruitment target of 3640 patients in three arms: best medical
treatment, CEA in addition to best medical treatment, and CAS in addition to best
medical treatment. The trial investigators recently redesigned the SPACE-2 protocol
such that now patients are enrolled into SPACE-2a comparing best medical treatment
with CAS in addition to best medical treatment, and SPACE-2b comparing best medical
treatment with CEA in addition to best medical treatment [242]. Interestingly, CREST-2
(the second Carotid Revascularization Endarterectomy versus Stenting Trial) has a
similar design and inclusion criteria to the new SPACE-2 protocol, in that neither trial
directly compares CAS with CEA. This role is taken by the second Asymptomatic
Carotid Surgery Trial (ACST-2) [174], which aims to compare the safety and efficacy of
CAS with CEA in 5000 patients with asymptomatic carotid stenosis who are “though to
need some procedural intervention” to prevent future stroke [243]. Recently-published
interim results from this study indicate that the risk of stroke, myocardial infarction or
death within 30 days of either procedure was low at 1.0% in 691 patients [174].
When clinicians are uncertain about which treatment will provide the greatest benefit to
the patient in return for an acceptably low risk of harm, it is appropriate to discuss with
them enrolment in a clinical trial. Patients can benefit from inclusion in randomized
166
clinical trials, even when allocated to the control therapy – close attention to control of
vascular risk factors and frequent follow-up may improve management of the patient’s
condition over and above routine clinical care. Almost all patients with low-risk carotid
stenosis can therefore be enrolled in one of the above trials, provided they can be
treated at a participating centre. Patients with low-risk symptomatic carotid stenosis
could be suitable for enrolment in ECST-2. When the medical team are uncertain as to
whether intervention is warranted in an asymptomatic patient they may enrol in ECST-
2, SPACE-2 or CREST-2. And when intervention is thought necessary in asymptomatic
patients, ACST-2 compares CAS with CEA for the long-term primary prevention of
stroke.
When the pace of change in stroke prevention is fast, there is a need for good-quality
evidence from randomized clinical trials to inform decision making from the individual
patient level up to the design of healthcare systems. If these four trials can recruit their
target numbers and report definitive results, then we may be some way further to
answering the question of “who should receive what treatment” for carotid stenosis.
167
21. Bibliography
[1] A. D. Lopez, C. D. Mathers, M. Ezzati, D. T. Jamison and C. J. L. Murray, “Global
and regional burden of disease and risk factors, 2001: systematic analysis of
population health data,” Lancet, vol. 367, no. 9524, pp. 1747-1757, 2006.
[2] WHO MONICA Project Principal Investigators, “The World Health Organization
MONICA Project (Monitoring Trends and Determinants in Cardiovascular
Disease): a major international collaboration,” J Clin Epidemiol, vol. 41, no. 2, pp.
105-114, 1998.
[3] The American Heart Association / American Stroke Association, “Definition and
evaluation of transient ischemic attack,” Stroke, vol. 40, no. 6, pp. 2276-2293,
2009.
[4] American Heart Association Statistics Committee and Stroke Statistics
Subcommittee, “Heart disease and stroke statistics - 2014 update: a report from
the American Heart Association,” Circulation, vol. 129, no. 3, pp. e28-e292, 2014.
[5] British Heart Foundation Health Promotion Research Group, “Coronary heart
disease statistics: a compendium of health statistics,” British Heart Foundation,
London, 2012.
[6] A. Hartmann, T. Rundek, H. Mast, R. L. Sacco, M. C. Paik and B. Boden-Albala,
“Mortality and causes of death after first ischemic stroke: The Northern
Manhattan Stroke Study,” Neurology, vol. 57, no. 11, pp. 2000-2005, 2001.
[7] The Stroke Unit Trialists' Collaboration, “Organised inpatient (stroke unit) care for
stroke (review),” Cochrane Library, no. 4, 2007.
[8] D. Doig and M. M. Brown, “Acute stroke care and management of carotid artery
stenosis,” in Cardiovascular disease and health in the older patient, Wiley-
Blackwell, 2013, pp. 261-298.
168
[9] K.-T. Khaw, “Epidemiology of stroke,” J Neurol Neurosurg Psych, vol. 61, no. 4,
pp. 333-338, 1996.
[10] V. L. Feigin, C. M. M. Lawes, D. A. Bennett and C. S. Anderson, “Stroke
epidemiology: a review of population-based studies of incidence, prevalence, and
case-fatality in the late 20th century,” Lancet Neurol, vol. 2, no. 1, pp. 43-53,
2003.
[11] D. P. Faxon, V. Fuster, P. Libby, J. A. Beckman, W. R. Hiatt and R. W.
Thompson, “Atherosclerotic vascular disease conference - writing group III:
pathophysiology,” Circulation, vol. 109, pp. 2617-2625, 2004.
[12] J. N. E. Redgrave, J. K. Lovett, P. J. Gallagher and P. M. Rothwell, “Histological
assessment of 526 symptomatic carotid plaques in relation to the nature and
timing of ischemic symptoms,” Circulation, vol. 113, pp. 2320-2328, 2006.
[13] T. Rundek, H. Arif, B. Boden-Albala, M. S. Elkind, M. C. Paik and R. L. Sacco,
“Carotid plaque, a subclinical precursor of vascular events: The Northern
Manhattan Study,” Neurology, vol. 70, no. 14, pp. 1200-1207, 2008.
[14] S. Prabhakaran, T. Rundek, R. Ramas, M. S. V. Elkind, M. C. Paik and B. Boden-
Albala, “Carotid plaque surface irregularity predicts ischemic stroke: The Northern
Manhattan Study,” Stroke, vol. 37, pp. 2696-2701, 2006.
[15] M. de Weerd, J. P. Greving, B. Hedblad, M. W. Lorenz, E. B. Mathiesen and D.
H. O'Leary, “Prevalence of asymptomatic carotid artery stenosis in the general
population: an individual participant meta-analysis,” Stroke, vol. 41, pp. 1294-
1297, 2010.
[16] B. Ahmed and H. Al-Khaffaf, “Prevalence of significant asymptomatic carotid
artery disease in patients with peripheral vascular disease: a meta-analysis,” Eur
J Vasc Endovasc Surg, vol. 37, no. 3, pp. 262-271, 2009.
[17] A. Steinvil, B. Sadeh, Y. Arbel, D. Justo, A. Belei and N. Borenstein, “Prevalence
and predictors of concomitant carotid and coronary artery atherosclerotic
disease,” J Am Coll Cardiol, vol. 57, no. 7, pp. 779-783, 2011.
169
[18] J. S. Fine-Edelstein, P. A. Wolf, D. H. O'Leary, H. Poehlman, A. J. Belanger and
C. S. Kase, “Precursors of extracranial carotid atherosclerosis in the Framingham
Study,” Neurology, vol. 44, no. 6, pp. 1046-1050, 1994.
[19] L. Marquardt, O. C. Geraghty, Z. Mehta and P. M. Rothwell, “Low risk of
ipsilateral stroke in patients with asymptomatic carotid stenosis on best medical
treatment: a propsective, population-based study,” Stroke, vol. 41, no. 1, pp. e11-
e17, 2010.
[20] J. K. Lovett, A. J. Coull and P. M. Rothwell, “Early risk of recurrence by subtype
of ischemic stroke in population-based incidence studies,” Neurology, vol. 62, no.
4, pp. 569-573, 2004.
[21] International Stroke Trial Collaborative Group, “The International Stroke Trial: A
randomised trial of aspirin, subcutaneous heparin, both, or neither among 19435
patients with acute ischaemic stroke,” Lancet, vol. 349, pp. 1569-1581, 1997.
[22] The ESPRIT Study Group, “Aspirin plus dipyridamole versus aspirin alone after
cerebral ischaemia of arterial origin (ESPRIT): randomised controlled trial,”
Lancet, vol. 367, pp. 1665-1673, 2006.
[23] The ProFESS Study Group, “Aspirin and extended-release dipyridamole versus
clopidogrel for recurrent stroke,” N Engl J Med, vol. 359, pp. 1238-1251, 2008.
[24] N. Chapman, R. Huxley, C. Anderson, M. G. Bousser, J. Chalmers and S.
Colman, “Effects of a perindopril-based blood pressure-lowering regimen on the
risk of recurrent stroke according to stroke subtype and medical history: the
PROGRESS trial,” Stroke, vol. 35, pp. 116-121, 2004.
[25] The Stroke Prevention by Aggressive Reduction in Cholesterol Levels (SPARCL)
Investigators, “High-dose atorvastatin after stroke or transient ischaemic attack,”
N Engl J Med, vol. 355, pp. 549-559, 2006.
[26] NHS Choices, “Stroke - prevention,” [Online]. Available:
http://www.nhs.uk/Conditions/Stroke/Pages/Prevention.aspx. [Accessed 8th July
2014].
170
[27] G. J. Hankey, C. P. Warlow and R. J. Sellar, “Cerebral angiographic risk in mild
cerebrovascular disease,” Stroke, vol. 21, pp. 209-222, 1990.
[28] P. S. Sidhu, “Ultrasound of the carotid and vertebral arteries,” Br Med Bulletin,
vol. 56, no. 2, pp. 346-366, 2000.
[29] J. M. Wardlaw, F. M. Chappell, J. J. K. Best, K. Wartolowska and E. Berry, “Non-
invasive imaging compared with intra-arterial angiography in the diagnosis of
symptomatic carotid stenosis: a meta-analysis,” Lancet, vol. 367, no. 9521, pp.
1503-1512, 2006.
[30] N. Magarelli, T. Scarabino, A. L. Simeone, F. Florio, A. Carriero and U. Salvolini,
“Carotid stenosis: a comparison between MR and spiral CT angiography,”
Neuroradiology, vol. 40, no. 6, pp. 367-373, 1998.
[31] M. E. DeBakey, “Successful carotid endarterectomy for cerebrovascular
insufficiency: nineteen-year follow-up,” JAMA, vol. 233, no. 10, pp. 1083-1085,
1975.
[32] H. H. G. Eastcott, G. W. Pickering and C. G. Rob, “Reconstruction of internal
carotid artery in a patient with intermittent attacks of hemiplegia,” Lancet, vol.
264, no. 6846, pp. 994-996, 1954.
[33] H. J. M. Barnett, F. Plum and J. N. Walton, “Carotid Endarterectomy — An
Expression of Concern,” Stroke, vol. 15, no. 6, pp. 941-943, 1984.
[34] N. C. Fode, T. M. Sundt, J. T. Robertson, S. J. Peerless and C. B. Shields,
“Multicenter retrospective review of results and complications of carotid
endarterectomy in 1981,” Stroke, vol. 17, no. 3, pp. 370-376, 1986.
[35] The North American Symptomatic Carotid Endarterectomy Trial Collaborators,
“Benefit of carotid endarterectomy in patients with symptomatic moderate or
severe stenosis,” N Engl J Med, vol. 339, pp. 1415-1425, 1998.
[36] The European Carotid Surgery Trialists' Collaborative Group, “Randomised trial
of endarterectomy for recently symptomatic carotid stenosis: final results of the
MRC European Carotid Surgery Trial (ECST),” Lancet, vol. 351, no. 9113, pp.
171
1379-1387, 1998.
[37] North American Symptomatic Carotid Endarterectomy Trial Collaborators,
“Beneficial effect of carotid endarterectomy in symptomatic patients with high-
grade carotid stenosis,” N Engl J Med, vol. 325, no. 7, pp. 445-453, 1991.
[38] The Carotid Endarterectomy Trialists' Collaboration, “Analysis of pooled data from
the randomised controlled trials of endarterectomy for symptomatic carotid
stenosis,” Lancet, vol. 361, no. 9352, pp. 107-116, 2003.
[39] M. R. Mayberg, S. E. Wilson, F. Yatsu, D. G. Weiss, L. Messina and L. A.
Hershey, “Carotid Endarterectomy and prevention of cerebral ischemia in
symptomatic carotid stenosis,” JAMA, vol. 266, no. 23, pp. 3289-3294, 1991.
[40] A. J. Fox, “How to measure carotid stenosis,” Radiology, vol. 186, no. 2, pp. 316-
318, 1993.
[41] A. J. Fox, M. Eliasziw and P. M. Rothwell, “Identification, prognosis, and
management of patients with carotid artery near occlusion,” Am J Neuroradiol,
vol. 26, pp. 2086-2094, 2005.
[42] D. Doig and M. M. Brown, “Carotid stenting versus endarterectomy,” Ann Rev
Med, vol. 63, pp. 259-276, 2012.
[43] The Carotid Endarterectomy Trialists' Collaboration, “Endarterectomy for
symptomatic carotid stenosis in relation to clinical subgroups and timing of
surgery.,” Lancet, vol. 363, no. 9413, pp. 915-924, 2004.
[44] The American College of Cardiology Foundation/American Heart Association
Task Force on Practice Guidelines,
“ASA/ACCF/AHA/AANN/AANS/ACR/ASNR/CNS/SAIP/SCAI/SIR/SNIS/SVM/SVS
guideline on the management of patients with extracranial carotid and vertebral
artery disease,” J Am Coll Cardiol, vol. 57, no. 8, pp. e16-e94, 2011.
[45] J. J. Ricotta, A. AbuRahma, E. Ascher, M. Eskandari, P. Faries and B. K. Lal,
“Updated Society for Vascular Surgery guidelines for management of extracranial
172
carotid disease,” J Vasc Surg, vol. 54, no. 3, pp. e1-e31, 2011.
[46] The ESVS Guidelines Collaborators, “ESVS guidelines. Invasive treatment for
carotid stenosis: indications, techniques,” Eur J Vasc Endovasc Surg, vol. 37, no.
4 (Suppl.), pp. 1-19, 2009.
[47] E. W. Massey, A. Heyman, C. Utley, C. Haynes and J. Fuchs, “Cranial nerve
paralysis following carotid endarterectomy,” Stroke, vol. 15, pp. 157-159, 1984.
[48] A. J. Maniglia and P. D. Han, “Cranial nerve injuries following carotid
endarterectomy: an analysis of 336 procedures,” Head & Neck, vol. 13, no. 2, pp.
121-124, 1991.
[49] E. Ballotta, G. da Giau, L. Renon, S. Narne, M. Saladini and E. Abbruzzese,
“Cranial and cervical nerve injuries after carotid endarterectomy: a prospective
study,” Surgery, vol. 125, pp. 85-91, 1999.
[50] M. D. Schauber, L. J. Fontenelle, J. W. Solomon and T. L. Hanson, “Cranial /
cervical nerve dysfunction after carotid endarterectomy,” J Vasc Surg, vol. 25, no.
3, pp. 481-487, 1997.
[51] M. J. Press, M. R. Chassin, J. Wang, S. Tuhrim and E. A. Halm, “Predicting
medical and surgical complications of carotid endarterectomy,” Arch Intern Med,
vol. 166, no. 8, pp. 914-920, 2006.
[52] A. Oldag, S. Schreiber, S. Schreiber, H.-J. Heinze, F. Meyer and M. Weber, “Risk
of wound haematoma at carotid endarerectomy under dual antiplatelet therapy,”
Langenbecks Arch Surg, vol. 397, pp. 1275-1282, 2012.
[53] N. Al-Mubarak, G. S. Roubin, S. S. Iyer, C. R. Gomez, M. W. Liu and J. J. Vitek,
“Carotid stenting for severe radiation-induced extracranial carotid artery occlusive
disease,” J Endovasc Ther, vol. 7, no. 1, pp. 36-40, 2000.
[54] M. D. Walker, J. R. Marler, M. Goldstein, P. A. Grady, J. F. Toole and W. H.
Baker, “Endarterectomy for asymptomatic carotid artery stenosis,” J Am Med
Assoc, vol. 273, no. 18, pp. 1421-1428, 1995.
173
[55] The MRC Asymptomatic Carotid Surgery Trial (ACST) Collaborative Group,
“Prevention of disabling and fatal strokes by successful carotid endarterectomy in
patients without recent neurological symptoms: randomised controlled trial,”
Lancet, vol. 363, pp. 1491-1502, 2004.
[56] The Asymptomatic Carotid Surgery Trial (ACST) Collaborative Group, “10-year
stroke prevention after successful carotid endarterectomy for asymptomatic
stenosis (ACST-1): a multicentre randomised trial,” Lancet, vol. 376, no. 9746,
pp. 1074-1084, 2010.
[57] C. T. Dotter and M. P. Judkins, “Transluminal treatment of arteriosclerotic
obstruction: description of a new technic and a preliminary report of its
application,” Circulation, vol. 30, pp. 654-670, 1964.
[58] C. T. Dotter, “Transluminally-placed coilspring endarterial tube grafts: long-term
patency in canine popliteal artery,” Investigative Radiol, vol. 4, no. 5, pp. 329-332,
1969.
[59] M. M. Brown, P. Butler, J. Gibbs, M. Swash and J. Waterston, “Feasibility of
percutaneous transluminal angioplasty for carotid artery stenosis,” J Neurol
Neurosurg Psych, vol. 53, no. 3, pp. 238-243, 1990.
[60] M. M. Brown, “Surgery, angioplasty and interventional neuroradiology,” Curr Opin
Neurol Neurosurg, vol. 6, no. 1, pp. 66-73, 1993.
[61] A. Gil-Peralta, A. Mayol, J. R. G. Marcos, A. Gonzalez, J. Ruano and F. Boza,
“Percutaneous transluminal angioplasty of the symptomatic atherosclerotic
carotid arteries: results, complications, and follow-up,” Stroke, vol. 27, no. 12, pp.
2271-2273, 1996.
[62] M. H. Wholey, N. Al-Mubarek and M. H. Wholey, “Upadated review of the global
carotid artery stent registry,” Catheter Cardiovasc Interv, vol. 60, no. 2, pp. 259-
266, 2003.
[63] The EVA-3S Investigators, “Carotid angioplasty and stenting with and without
cerebral protection: clinical alert from the Endarterectomy Versus Angioplasty in
patients with Symptomatic Severe carotid Stenosis (EVA-3S) trial,” Stroke, vol.
174
35, no. 1, pp. e18-e20, 2004.
[64] The ICSS-MRI Substudy Investigators, “New ischaemic brain lesions on MRI
after stenting or endarterectomy for symptomatic severe carotid stenosis: a
substudy of the International Carotid Stenting Study (ICSS),” Lancet Neurol, vol.
9, no. 4, pp. 353-362, 2010.
[65] J. C. Parodi, L. M. Ferreira, G. Sicard, R. la Mura and S. Fernandez, “Cerebral
protection during carotid stenting using flow reversal.,” J Vasc Surg, vol. 41, no.
3, pp. 416-422, 2005.
[66] L. H. de Castro-Afonso, L. G. Abud, J. G. Rolo, A. C. dos Santos, L. de Oliveira
and C. M. A. Barreira, “Flow reversal versus filter protection: a pilot carotid artery
stenting trial,” Circ Cardiovasc Intervent, vol. 6, pp. 552-559, 2013.
[67] B. Reimers, N. Corvaja, S. Moshiri, S. Saccà, R. Albiero and C. dr Mario,
“Cerebral protection with filter devices during carotid artery stenting,” Circulation,
vol. 104, no. 1, pp. 12-15, 2001.
[68] E. B. Diethrich, “Indications for carotid artery stenting: a preview of the potential
derived from early clinical experience,” J Endovasc Surg, vol. 3, no. 2, pp. 132-
139, 1996.
[69] M. M. Taha, H. Sakaida, F. Asakura, M. Maeda, N. Toma and T. Sano, “Access
site complications with carotid angioplasty and stenting,” Surg Neurol, vol. 68, no.
4, pp. 431-437, 2007.
[70] A. R. Naylor, “Randomized study of carotid angioplasty and stenting versus
carotid endarterectomy: a stopped trial,” J Vasc Surg, vol. 28, no. 2, pp. 326-334,
1998.
[71] W. H. Brooks, R. R. McClure, M. R. Jones, T. L. Coleman and L. R. N. Breathitt,
“Carotid angioplasty and stenting versus carotid endarterectomy for the treatment
of asymptomatic carotid stenosis: a randomized trial in a community hospital,”
Neurosurgery, vol. 54, no. 2, pp. 318-325, 2004.
175
[72] A. Hoffman, S. Engelter, C. Taschner, A. Mendelowitsch, A. Merlo and E.-W.
Radue, “Carotid stenting versus endarterectomy - a prospective randomised
controlled single-centre trial with long-term follow-up (BACASS),” Schweizer Arch
Neurol Psych, vol. 159, pp. 84-89, 2008.
[73] M. J. Alberts, “Results of a multicenter prospective randomized trial of carotid
artery stenting vs. carotid endarterectomy,” Stroke, vol. 32, p. 325, 2001.
[74] J. S. Yadav, M. H. Wholey, R. E. Kuntz, P. Fayad, B. T. Katzen and G. J.
Mishkel, “Protected carotid-artery stenting versus endarterectomy in high-risk
patients,” N Engl J Med, vol. 351, no. 15, pp. 1493-1501, 2004.
[75] The CAVATAS Investigators, “Endovascular versus surgical treatment in patients
with carotid stenosis in the Carotid And Vertebral Artery Transluminal Angioplasty
Study (CAVATS): a randomized trial,” Lancet, vol. 357, no. 9270, pp. 1729-1737,
2001.
[76] The CAVATAS Investigators, “Endovascular treatment with angioplasty or
stenting versus endarterectomy in patients with carotid artery stenosis in the
Carotid And Vertebral Artery Transluminal Angioplasty Study (CAVATAS): long-
term follow-up of a randomised trial,” Lancet Neurol, vol. 8, no. 10, pp. 898-907,
2009.
[77] The SPACE Collaborative Group, “30 day results from the SPACE trial of stent-
protected angioplasty versus carotid endarterectomy in symptomatic patients: a
randomised non-inferiority trial,” Lancet, vol. 368, no. 9543, pp. 1239-1247, 2006.
[78] H.-H. Eckstein, P. Ringleb, J.-R. Allenberg, J. Berger, G. Fraedrich and W.
Hacke, “Results of the Stent-Protected Angioplasty versus Carotid
Endarterectomy (SPACE) study to treat symptomatic stenoses at 2 years: a
multinational, prospective, randomised trial,” Lancet Neurol, vol. 7, no. 10, pp.
893-902, 2008.
[79] The EVA-3S Investigators, “Endarterectomy versus stenting in patients with
symptomatic severe carotid stenosis,” N Engl J Med, vol. 355, pp. 1660-1671,
2006.
176
[80] The EVA-3S Investigators, “Endarterectomy Versus Angioplasty in patients with
Symptomatic Severe carotid Stenosis (EVA-3S) trial: results up to 4 years from a
randomised, multicentre trial,” Lancet Neurol, vol. 7, no. 10, pp. 885-892, 2008.
[81] The ICSS Investigators, “International Carotid Stenting Study: protocol for a
randomised clinical trial comparing carotid stenting with endarterectomy in
symptomatic carotid artery stenosis,” Cerebrovasc Dis, vol. 18, pp. 69-74, 2004.
[82] The ICSS Investigators, “Carotid artery stenting compared with endarterectomy in
patients with symptomatic carotid stenosis (International Carotid Stenting Study):
an interim analysis of a randomised controlled trial.,” Lancet, vol. 375, no. 9719,
pp. 985-997, 2010.
[83] M. M. Brown, J. Dobson, D. Doig, R. L. Featherstone and E. L. Turner, “Primary
analysis of the International Carotid Stenting Study: a randomised comparison of
the effectiveness of carotid stenting and endarterectomy in preventing long-term
stroke in patients with symptomatic carotid stenosis,” in European Stroke
Conference, Lisbon, 2012.
[84] The CREST Investigators, “Stenting versus endarterectomy for treatment of
carotid-artery stenosis,” N Engl J Med, vol. 363, no. 1, pp. 11-23, 2010.
[85] The CREST Investigators, “Safety of stenting and endarterectomy by
symptomatic status in the Carotid Revascularization Endarterectomy Versus
Stenting Trial (CREST),” Stroke, vol. 42, pp. 675-680, 2011.
[86] S. Bangalore, S. Kumar, J. Wetterslev, A. A. Bavry, C. Gluud and D. E. Cutlip,
“Carotid stenting versus carotid endarterectomy: meta-analysis and diversity-
adjusted trial sequential analysis of randomized trials,” Arch Neurol, vol. 68, no. 2,
pp. 172-184, 2010.
[87] F. Ling and Q. L. Jiao, “Preliminary report of trial of endarterectomy versus
stenting for the treatment of carotid atherosclerotic stenosis in China (TESCAS-
C),” Chin J Cerebrovasc Dis, vol. 3, no. 1, pp. 4-8, 2006.
[88] ClinicalTrials.gov, [Online]. Available:
177
http://clinicaltrials.gov/ct2/show/NCT00106938. [Accessed 7th May 2014].
[89] M. P. Janssen, G. J. de Borst, W. P. T. M. Mali, L. J. Kappelle, F. L. Moll and R.
G. A. Ackerstaff, “Carotid stenting versus carotid endarterectomy: evidence basis
and cost implications,” Eur J Vasc Endovasc Surg, vol. 36, no. 3, pp. 258-264,
2008.
[90] W. A. Gray, H. J. White, D. M. Barrett, G. Chandran, R. Turner and M. Reisman,
“Carotid stenting and endarterectomy: a clinical and cost comparison of
revascularization strategies,” Stroke, vol. 33, pp. 1063-1070, 2002.
[91] M. J. Donovan, D. E. Ramirez, G. D. Crenshaw, T. A. Smith, H. A. Bazan and W.
C. Sternbergh III, “Hospital reimbursement for carotid stenting and
endarterectomy,” J Endovasc Ther, vol. 21, no. 2, pp. 296-302, 2014.
[92] M. A. Almekhlafi, M. D. Hill, S. Wiebe, M. Goyal, D. Yavin and J. H. Wong, “When
is carotid angioplasty and stenting the cost-effective alternative for
revascularization of symptomatic carotid stenosis? A Canadian health system
perspective,” ANJR, vol. 35, pp. 327-332, 2014.
[93] The Safe Surgery Saves Lives Study Group, “A surgical safety checklist to
reduce morbidity and mortality in a global population,” New Engl J Med, vol. 360,
pp. 491-499, 2009.
[94] A. A. Gawande, E. J. Thomas, M. J. Zinner and T. A. Brennan, “The incidence
and nature of surgical adverse events in Colorado and Utah in 1992,” Surgery,
vol. 126, no. 1, pp. 66-75, 1999.
[95] A. K. Kable, R. W. Gibberd and A. D. Spigelman, “Adverse events in surgical
patients in Australia,” Int J Qual Health Care, vol. 14, no. 4, pp. 269-276, 2002.
[96] K. Rerkasem and P. M. Rothwell, “Temporal trends in the risk of stroke and death
due to endarterectomy for symptomatic carotid stenosis: an updated systematic
review,” Eur J Vasc Endovasc Surg, vol. 37, no. 5, pp. 504-511, 2009.
[97] S. D. Goode, T. J. Cleveland and P. A. Gaines, “United Kingdom Carotid Artery
Stent Registry: short- and long-term outcomes,” Cardiovasc Intervent Radiol, vol.
178
36, pp. 1221-1231, 2013.
[98] P. M. Rothwell and C. P. Warlow, “Prediction of benefit from carotid
endarterectomy in individual patients: a risk-modelling study,” Lancet, vol. 353,
no. 9170, pp. 2105-2110, 1999.
[99] M. Frego, A. Bridda, C. Ruffolo, M. Scarpa, L. Polese and G. Bianchera, “The
hostile neck does not increase the risk of carotid endarterectomy,” J Vasc Surg,
vol. 50, no. 1, pp. 40-47, 2009.
[100] A. P. Gasparis, L. Ricottaa, S. A. Cuadra, D. J. Char, W. A. Purtill and P. S. van
Bemmelen, “High-risk carotid endarterectomy: fact or fiction,” J Vasc Surg, vol.
37, no. 1, pp. 40-46, 2003.
[101] J. S. Yadav, “Carotid stenting in high-risk patients: design and rationale of the
SAPPHIRE trial,” Clev Clin J Med, vol. 71, pp. S45-S46, 2004.
[102] G. Mozes, T. M. Sullivan, D. R. Torres-Russotto, T. C. Bower, T. L. Hoskin and S.
M. Sampaio, “Carotid endarterectomy in SAPPHIRE-eligible high-risk patients:
implications for selecting patients for carotid angioplasty and stenting,” J Vasc
Surg, vol. 39, no. 5, pp. 958-965, 2004.
[103] K. Rerkasem and P. M. Rothwell, “Local versus general anesthetic for carotid
endarterectomy.,” Stroke, vol. 40, pp. e584-e585, 2009.
[104] The GALA Trial Collaborative Group, “General anaesthesia versus local
anaesthesia for carotid surgery (GALA): a multicentre, randomised controlled
trial.,” Lancet, vol. 372, no. 9656, pp. 2132-2142, 2008.
[105] P. M. Rothwell, M. Eliasziw and S. A. Gutnikov, “Endarterectomy for symptomatic
carotid stenosis in relation to clinical subgroups and timing of surgery,” Lancet,
vol. 363, pp. 915-924, 2004.
[106] K. Rerkasem and P. M. Rothwell, “Routine or selective carotid artery shunting for
carotid endarterectomy (and different methods of monitoring in selective
shunting),” Stroke, vol. 41, pp. e53-e54, 2010.
179
[107] K. Rerkasem and P. M. Rothwell, “Patch angioplasty versus primary closure for
carotid endarterectomy (review),” Cochrane Library, 2009.
[108] D. A. Kiskinis, N. A. Saratzis and A. Saratzis, “Eversion carotid endarterectomy
technique,” in European Manual of Medicine - Vascular surgery, Springer, 2007,
pp. 155-160.
[109] T. S. Riles, A. M. Imparato, G. R. Jacobowitz, P. J. Lamparello, G. Giangola and
M. A. Adelman, “The cause of perioperative stroke after carotid endarterectomy,”
J Vasc Surg, vol. 19, no. 2, pp. 206-216, 1994.
[110] J. A. Cowan Jr, J. B. Dimick, B. G. Thompson, J. C. Stanley and G. R. Upchurch
Jr, “Surgeon volume as an indicator of outcomes after carotid endarterectomy: an
effect independent of specialty practice and hospital volume,” J Am Coll Surg, vol.
195, no. 6, pp. 814-821, 2002.
[111] D. S. Kucey, B. Bowyer, K. Iron, P. Austin, G. Anderson and J. V. Tu,
“Determinants of outcome after carotid endarterectomy,” J Vasc Surg, vol. 28, no.
6, pp. 1051-1058, 1998.
[112] S. T. Ruby, D. Robinson, J. T. Lynch and H. Mark, “Outcome analysis of carotid
endarterectomy in Connecticut: the impact of volume and specialty,” Ann Vasc
Surg, vol. 10, no. 1, pp. 22-26, 1996.
[113] P. J. E. Holt, J. D. Poloniecki, I. M. Loftus and M. M. Thompson, “Meta-analysis
and systematic review of the relationship between hospital volume and outcome
following carotid endarterectomy,” Eur J Vasc Endovasc Surg, vol. 33, no. 6, pp.
645-651, 2007.
[114] The Royal College of Surgeons of England, “Reshaping surgical services:
principles for change,” 2013.
[115] The Carotid Stenting Trialists' Collaboration, “Short-term outcome after stenting
versus endarterectomy for symptomatic carotid stenosis: a preplanned meta-
analysis of individual patient data,” Lancet, vol. 376, no. 9746, pp. 1062-1073,
2010.
180
[116] The CREST Investigators, “Age and outcomes after carotid stenting and
endarterectomy: The Carotid Revascularization Endarterectomy versus Stenting
Trial,” Stroke, vol. 42, pp. 3484-3490, 2011.
[117] J. Ederle, I. Davagnanam, H. B. van der Worp, G. S. Venables, P. A. Lyrer and R.
L. Featherstone, “Effect of white-matter lesions on the risk of periprocedural
stroke after carotid artery stenting versus endarterectomy in the International
Carotid Stenting Study (ICSS): a prespecified analysis of data from a randomised
trial,” Neurology, vol. 12, no. 9, pp. 866-872, 2013.
[118] M. M. Brown and R. Raine, “Should sex influence the choice between carotid
stenting and carotid endarterectomy?,” Lancet Neurol, vol. 10, no. 6, pp. 494-496,
2011.
[119] J. P. Hart, P. Peeters, J. Verbist, K. Deloose and M. Bosiers, “Do device
characteristics impact outcome in carotid artery stenting?,” J Vasc Surg, vol. 44,
no. 4, pp. 725-730, 2006.
[120] The EVA-3S Investigators, “Anatomical and technical factors associated with
stroke or death during carotid angioplasty and stenting: results from the
Endarterectomy Versus Angioplasty in patients with Symptomatic Severe carotid
Stenosis (EVA-3S) trial and systematic review,” Stroke, vol. 42, pp. 380-388,
2011.
[121] S. Sayeed, S. F. Stanziale, M. H. Wholey and M. S. Makaroun, “Angiographic
lesion characteristics can predict adverse outcomes after carotid artery stenting,”
J Vasc Surg, vol. 47, no. 1, pp. 81-87, 2008.
[122] C. Setacci, E. Chisci, F. Setacci, F. Iacoponi, G. de Donato and A. Rossi, “Siena
carotid artery stenting score. A risk modelling study for individual patients.,”
Stroke, vol. 41, pp. 1259-1265, 2010.
[123] S. Macdonald, R. Lee, R. Williams and G. Stansby, “Towards safer carotid artery
stenting: a scoring system for anatomic suitability,” Stroke, vol. 40, pp. 1698-
1703, 2009.
181
[124] P. M. Rothwell and C. P. Warlow, “Interpretation of operative risks of individual
surgeons,” Lancet, vol. 353, no. 9161, p. 1325, 1999.
[125] J. Fiehler, S. J. Bakke, A. Clifton, E. Houdart, O. Jansen and D. Rüfenacht, “Plea
of the defence - critical comments on the interpretation of EVA3S, SPACE and
ICSS,” Neuroradiology, vol. 52, pp. 601-610, 2010.
[126] D. Calvet, A. Algra, J.-P. Becquemin, L. H. Bonati, J. Dobson and G. Fraedrich,
“Is there an operator effect? A pooled analysis from the Carotid Stenting Trialists'
Collaboration,” Stroke, vol. 45, pp. 00-00, 2014.
[127] W. N. Kernan, B. Ovbiagele, H. R. Black, D. M. Bravata, M. I. Chimowitz and M.
D. Ezekowitz, “Guidelines for the prevention of stroke in patients with stroke and
transient ischemic attack: a guideline for healthcare professionals from the
American Heart Association / American Stroke Association,” 2014. [Online].
Available: doi:10.1161/STR.0000000000000024. [Accessed 8th May 2014].
[128] NICE, “Stroke: diagnosis and initial management of acute stroke and transient
ischaemic attack (TIA),” NICE, 2008.
[129] NICE, “Carotid artery stent placement for symptomatic extracranial carotid
stenosis,” NICE, 2011.
[130] J. D. Kakisis, E. D. Avgerinos, C. N. Antonopoulos, T. G. Giannakopoulos, K.
Moulakakis and C. D. Liapis, “The European Society for Vascular Surgery
guidelines for carotid intervention: an updated independent assessment and
literature review,” Eur J Vasc Endovasc Surg, vol. 44, no. 3, pp. 238-243, 2012.
[131] A. H. Y. Lee, J. Busby, M. Brooks and W. Hollingworth, “Uptake of carotid artery
stenting in England and subsequent vascular admissions: an appropriate
response to emerging evidence?,” European Journal of Vascular and
Endovascular Surgery, vol. 46, no. 3, pp. 282-289, 2013.
[132] J. S. McDonald, R. J. McDonald, J. Fan, G. Lanzino, D. F. Kallmes and H. J.
Kloft, “Effect of CREST Findings on Carotid Revascularization Practice in the
United States,” Stroke and Cerebrovasc Dis, p. (online), 2015.
182
[133] K. Rerkasem and P. M. Rothwell, “Systematic review of randomized controlled
trials of carotid endarterectomy for symptomatic carotid stenosis,” Stroke, vol. 42,
pp. e543-e544, 2011.
[134] R. Bond, K. Rerkasem, C. P. Shearman and P. M. Rothwell, “Time trends in the
published risks of stroke and death due to endarterectomy for symptomatic
carotid stenosis.,” Cerebrovasc Dis, vol. 18, pp. 37-46, 2004.
[135] J. P. Simons, P. P. Goodney, D. T. Baril, B. W. Nolan, N. D. Hevelone and J. L.
Cronenwett, “The effect of postoperative stroke and myocardial infarction on long-
term survival after carotid revascularization,” J Vasc Surg, vol. 57, no. 6, pp. 1-8,
2013.
[136] N. Rudarakanchana, A. W. Halliday, D. Kamugasha, R. Grant, S. Waton and M.
Horrocks, “Current practice of carotid endarterectomy in the UK,” Brit J Surg, vol.
99, no. 2, pp. 209-216, 2012.
[137] The GALA Trial Collaborators, “Carotid endarterectomy: technical practices of
surgeons participating in the GALA trial,” Eur J Vasc Endovasc Surg, vol. 36, no.
4, pp. 385-389, 2008.
[138] R. Bond, K. Rerkasem, R. Cuffe and P. M. Rothwell, “A systematic review of the
associations between age and sex and the operative risks of carotid
endarterectomy,” Cerebrovasc Dis, vol. 20, pp. 69-77, 2005.
[139] J. E. Frawley and R. G. Hicks, “Risk factors for peri-operative stroke complicating
carotid endarterectomy: selective analysis of a prospective audit of 1000
consecutive operations,” Aust NZ J Surg, vol. 70, no. 1, pp. 52-56, 2000.
[140] C. B. Rockman, K. Garg, G. R. Jacobowitz, J. S. Berger, F. F. Mussa and N. S.
Cayne, “Outcome of carotid artery interventions among female patients, 2004 to
2005,” J Vasc Surg, vol. 53, no. 6, pp. 1457-1464, 2011.
[141] J. Jim, E. D. Dillavou, G. R. Upchurch, N. H. Osborne, C. T. Kenwood and F. S.
Siami, “Gender-specific 30-day outcomes after carotid endarterectomy and
carotid artery stenting in the Society for Vascular Surgery Vascular Registry,” J
183
Vasc Surg, vol. 59, no. 3, pp. 742-748, 2014.
[142] J. Krejza, M. Arkuszewski, S. E. Kasner, J. Weigele, A. Ustymowicz and R. W.
Hurst, “Carotid artery diameter in men and women and the relation to body and
neck size,” Stroke, vol. 37, pp. 1103-1105, 2006.
[143] J. Golledge, R. Cuming, D. K. Beattie, A. H. Davies and R. M. Greenhalgh,
“Influence of patient-related variables on the outcome of carotid endarterectomy,”
J Vasc Surg, vol. 24, no. 1, pp. 120-126, 1996.
[144] M. K. Kapral, H. Wang, P. C. Austin, J. Fang, D. Kucey, B. Bowyer and J. V. Tu,
“Sex Differences in Carotid Endarterectomy Outcomes: Results From the Ontario
Carotid Endarterectomy Registry,” Stroke, vol. 34, pp. 1124-1125, 2003.
[145] D. M. Laman, G. H. Wieneke, H. van Duijn and A. C. van Huffelen, “High embolic
rate early after carotid endarterectomy is associated with early cerebrovascular
complications, especially in women.,” J Vasc Surg, vol. 36, pp. 278-284, 2002.
[146] A. G. den Hartog, A. Algra, F. L. Moll and G. J. de Borst, “Mechanisms of gender-
related outcome differences after carotid endarterectomy,” J Vasc Surg, vol. 52,
pp. 1062-1071, 2010.
[147] P. M. Rothwell, M. Eliasziw, S. A. Gutnikov, C. P. Warlow and H. J. Barnett, “Sex
difference in the effect of time from symptoms to surgery on benefit from carotid
endarterectomy for transient ischemic attack and nondisabling stroke,” Stroke,
vol. 35, pp. 2855-2861, 2004.
[148] S. Strömberg, J. Gelin, T. Österberg, G. M. Bergström, L. Karlström and K.
Österberg, “Very urgent carotid endarterectomy confers increased procedural
risk,” Stroke, vol. 43, pp. 1331-1335, 2012.
[149] P. M. Rothwell, J. Slattery and C. P. Warlow, “Clinical and angiographic
predictors of stroke and death from carotid endarterectomy: systematic review,”
BMJ, vol. 315, pp. 1571-1577, 1997.
[150] A. Altinbas, A. Algra, M. M. Brown, R. L. Featherstone, L. J. Kappelle and G. J.
de Borst, “Effects of carotid endarterectomy or stenting on haemodynamic
184
complications in the International Carotid Stenting Study: a randomized
comparison,” Int J Stroke, vol. 9, no. 3, pp. 284-290, 2013.
[151] J. H. Wong, J. M. Findlay and M. E. Suarez-Almazor, “Haemodynamic instability
after carotid endarterectomy: risk factors and associations with operative
complications,” Neurosurgery, vol. 41, no. 1, pp. 35-41, 1997.
[152] B. Towne and V. M. Bernhard, “The relationship of postoperative hypertension to
complications following carotid endarterectomy,” Surgery, vol. 88, no. 4, pp. 575-
580, 1980.
[153] C. D. Liapis, P. R. F. Sir Bell and D. Mikhailidis, “ESVS Guidelines. Invasive
treatment for carotid stenosis: indications, techniques,” Eur J Vasc Endovasc, vol.
37, pp. 1-19, 2009.
[154] B. Rantner, H.-H. Eckstein, P. Ringleb, K. D. Woelfle, H. Bruijnen and C.
Schmidauer, “American Society of Anaesthesiology and Rankin as predictive
parameters for the outcome of carotid endarterectomy within 28 days after an
ischemic stroke,” J Stroke Cerebrovasc Dis, vol. 15, no. 3, pp. 114-120, 2006.
[155] J. V. Tu, H. Wang, B. Bowyer, L. Green, J. Fang and D. Kucey, “Risk factors for
death or stroke after carotid endarterectomy: observations from the Ontario
Carotid Endarterectomy Registry,” Stroke, vol. 34, pp. 2568-2573, 2003.
[156] G. Kuhan, E. D. Gardiner, A. F. Abidia, I. C. Chetter, P. M. Renwick and B. F.
Johnson, “Risk modelling study for carotid endarterectomy,” Brit J Surg, vol. 88,
no. 12, pp. 1590-1594, 2001.
[157] S. K. Kakkos, A. Nicolaides, M. Griffin, M. Sabetai, S. Dhanjil and D. J. Thomas,
“Factors associated with mortality in patients with asymptomatic carotid stenosis:
results from the ACSRS Study,” Int Angiol, vol. 24, no. 3, pp. 221-230, 2005.
[158] D. C. McCrory, L. B. Goldstein, G. P. Samsa, E. Z. Oddone, P. B. Landsman and
W. S. Moore, “Predicting complications of carotid endarterectomy,” Stroke, vol.
24, pp. 1285-1291, 1993.
185
[159] K. Bekelis, S. F. Bakhoum, A. Desai, T. A. Mackenzie, P. Goodney and N.
Labropoulos, “A risk factor-based predictive model of outcomes in carotid
endarterectomy: the National Surgical Quality Improvement Program 2005-2010,”
Stroke , vol. 44, pp. 1085-1090, 2013.
[160] E. A. Halm, S. Tuhrim, J. J. Wang, C. Rockman, T. S. Riles and M. R. Chassin,
“Risk factors for perioperative death and stroke after carotid endarterectomy:
results of the New York Carotid Artery Surgery Study,” Stroke, vol. 40, pp. 221-
229, 2009.
[161] P. P. Goodney, D. S. Likosky and J. L. Cronenwett, “Factors associated with
stroke or death after carotid endarterectomy in Northern New England,” J Vasc
Surg, vol. 48, no. 5, pp. 1139-1145, 2008.
[162] L. Calvillo-King, L. Xuan, S. Zhang, S. Tuhrim and E. H. Halm, “Predicting risk of
perioperative death and stroke after carotid endarterectomy in asymptomatic
patients: derivation and validation of a clinical risk score,” Stroke, vol. 41, no. 12,
pp. 2786-2794, 2010.
[163] T. F. Kresowik, D. Bratzler, H. R. Karp, R. A. Hemann, M. E. Hendel and S. L.
Grund, “Multistate utilization, processes, and outcomes of carotid
endarterectomy,” J Vasc Surg, vol. 33, no. 2, pp. 227-235, 2001.
[164] E. A. Halm, E. L. Hannan, M. Rojas, S. Tuhrim, T. S. Riles and C. B. Rockman,
“Clinical and operative predictors of outcomes of carotid endarterectomy,” J Vasc
Surg, vol. 42, no. 3, pp. 420-428, 2005.
[165] B. Kragsterman, K. Logason, A. Ahari, T. Troëng, H. Parsson and D. Bergqvist,
“Risk factors for complications after carotid endarterectomy - a population-based
study,” Eur J Vasc Endovasc Surg, vol. 28, no. 1, pp. 98-103, 2004.
[166] D. G. Altman, Y. Vergouwe, P. Royston and K. G. M. Moons, “Prognosis and
prognostic research: validating a prognostic model,” BMJ, vol. 338, pp. 1432-
1435, 2009.
[167] D. G. Altman and P. Royston, “What do we man by validating a prognostic
186
model?,” Stat Med, vol. 19, pp. 453-473, 2000.
[168] A. F. AbuRahma, P. A. Stone, M. Srivastava, S. M. Hass, A. Y. Mousa and L. S.
Dean, “The effect of surgeon's specialty and volumeon the perioperative outcome
of carotid endarterectomy,” J Vasc Surg, vol. 58, no. 3, pp. 666-672, 2013.
[169] P. M. Rothwell, Z. Mehta, S. C. Howard, S. A. Gutnikov and C. P. Warlow, “From
subgroups to individuals: general principles and the example of carotid
endarterectomy,” Lancet, vol. 365, no. 9455, pp. 256-265, 2005.
[170] A. Barth and C. Bassetti, “Patient selection for carotid endarterectomy: how far is
risk modelling applicable to the individual?,” Stroke, vol. 34, no. 2, pp. 524-527,
2003.
[171] E. Touzé, L. Trinquart, R. Felgueiras, K. Rerkasem, L. H. Bonati and G.
Meliksetyan, “A clinical rule (Sex, Contralateral occlusion, Age, and Restenosis)
to select patients for stenting versus carotid endarterectomy: systematic review of
observational studies with validation in randomized trials,” Stroke, vol. 44, no. 12,
pp. 3394-3400, 2013.
[172] F. Kennedy, R. L. Featherstone and M. M. Brown, “Aktueller Status von ECST-2,”
Gefässchirurgie, vol. 18, no. 7, pp. 633-637, 2013.
[173] The SPACE-2 Study Group, “Stent-protected angioplasty in asymptomatic carotid
artery stenosis vs endarterectomy: SPACE-2 - a three-arm randomised-controlled
clinical trial,” Int J Stroke, vol. 4, no. 4, pp. 294-299, 2009.
[174] The ACST-2 Collaborative Group, “Status Update and Interim Results from the
Asymptomatic Carotid Surgery Trial-2 (ACST-2),” Eur J Vasc Endovasc Surg, vol.
46, no. 5, pp. 510-518, 2013.
[175] The North American Symptomatic Carotid Endarterectomy Trial (NASCET)
Collaborators, “The North American Symptomatic Carotid Endarterectomy Trial:
surgical results in 1415 patients,” Stroke, vol. 30, no. 9, pp. 1751-1758, 1999.
[176] L. H. Bonati, P. Lyrer, J. Ederle, R. Featherstone and M. M. Brown,
“Percutaneous transluminal balloon angioplasty and stenting for carotid artery
187
stenosis,” The Cochrane Library, no. 9, 2012.
[177] A. J. Greenstein, M. R. Chassin, J. Wang, C. B. Rockman, T. S. Riles and S.
Tuhrim, “Association between minor and major surgical complications after
carotid endarterectomy: results of the New York Carotid Artery Surgery study.,” J
Vasc Surg, vol. 46, no. 6, pp. 1138-1146, 2007.
[178] D. D. Self, G. L. Bryson and P. J. Sullivan, “Risk factors for post-carotid
endarterectomy haematoma formation,” Can J Anesth, vol. 46, no. 7, pp. 365-
640, 1999.
[179] E. J. Cunningham, R. Bond, M. R. Mayberg, C. P. Warlow and P. M. Rothwell,
“Risk of persistent cranial nerve injury after carotid endarterectomy,” J Neurosurg,
vol. 101, no. 3, pp. 445-448, 2004.
[180] M. Fokkema, G. J. de Borst, B. W. Nolan, J. Indes, D. B. Buck and R. C. Lo,
“Clinical relevance of cranial nerve injury following carotid endarterectomy,” Eur J
Vasc Endovasc Surg, vol. 47, no. 1, pp. 2-7, 2014.
[181] R. J. Hye, A. Mackey, M. D. Hill, J. H. Voeks, D. J. Cohen, K. Wang, M. Tom and
T. G. Brott, “Incidence, outcomes, and effect on quality of life of cranial nerve
injury in the Carotid Revascularization Endarterectomy versus Stenting Trial,” J
Vasc Surg, vol. in press, 2015.
[182] S. M. M. Gisbert, V. A. S. Almonacil, J. M. Z. García, B. G. Gascó, F. J. G.
Palonés and E. O. Monzón, “Predictors of cervical bleeding after carotid
endarterectomy,” Ann Vasc Surg, vol. 28, no. 2, pp. 366-374, 2014.
[183] J. M. Kunkel, E. R. Gomez, M. J. Spebar, R. J. Delgado, B. S. Jarstfer and G. J.
Collins Jr, “Wound haematomas after carotid endarterectomy,” Am J Surg, vol.
148, no. 6, pp. 844-847, 1984.
[184] J. D. Beard, J. Mountney, J. M. Wilkinson, A. Payne, J. Dicks and D. Mitton,
“Prevention of postoperative wound haematomas and hyperperfusion following
carotid endarerectomy,” Eur J Vasc Endovasc Surg, vol. 21, no. 6, pp. 490-493,
2001.
188
[185] ESPRIT Study Group, “Aspirin plus dipyridamole versus aspirin alone after
cerebral ischaemia of arterial origin (ESPRIT): randomised controlled trial,”
Lancet, vol. 367, no. 9523, pp. 1665-1673, 2006.
[186] The ProFESS Study Group, “Aspirin and extended-release dipyridamole versus
clopidogrel for recurrent stroke,” New Engl J Med, vol. 359, no. 12, pp. 1238-
1251, 2008.
[187] The MATCH Investigators, “Aspirin and clopidogrel compared with clopidogrel
alone after recent ischaemic stroke or transient ischaemic attack in high-risk
patients (MATCH): randomised, double-blind, placebo-controlled trial,” Lancet,
vol. 364, no. 9431, pp. 331-337, 2004.
[188] J. Smout and G. Stansby, “Current practice in the use of antiplatelet agents in the
peri-operative period by UK vascular surgeons,” Ann R Coll Surg Engl, vol. 85,
no. 2, pp. 97-101, 2003.
[189] L. H. Bonati and G. Fraedrich, “Age modifies the relative risk of stenting versus
endarterectomy for symptomatic carotid stenosis - a pooled analysis of EVA-3S,
SPACE and ICSS,” Eur J Vasc Endovasc Surg, vol. 41, no. 2, pp. 153-158, 2011.
[190] O. Jansen, J. Fiehler, M. Hartmann and H. Brückmann, “Protection or
nonprotection in carotid stent angioplasty: the influence of interventional
techniques on outcome data from the SPACE trial,” Stroke, vol. 40, no. 3, pp.
841-846, 2009.
[191] M. Bosiers, G. de Donato and K. Deloose, “Does free cell area influence the
outcome in carotid artery stenting?,” Eur J Vasc Endovasc, vol. 33, no. 2, pp.
135-143, 2007.
[192] S. Schnaudigel, K. Gröschel, S. M. Pilgram and A. Kastrup, “New brain lesions
after carotid stenting versus carotid endarterectomy: a systematic review of the
literature,” Stroke, vol. 39, no. 6, pp. 1911-1919, 2008.
[193] C. Rogers and E. R. Edelman, “Endovascular Stent Design Dictates Experimental
Restenosis and Thrombosis,” Circulation, vol. 91, no. 12, pp. 2995 -3001, 1995.
189
[194] T. Obeid, D. J. Arnaoutakis, I. Ardhuidese, U. Qazi, C. J. Abularrage, J. Black, B.
Perler and M. Malas, “Poststent ballooning is associated with increased
periprocedural stroke and death rate in carotid artery stenting,” J Vasc Surg, p. In
press, 2015.
[195] A. Kastrup, K. Gröschel, H. Krapf, B. R. Brehm, J. Dichgans and J. Schulz, “Early
outcome of carotid angioplasty and stenting with and without cerebral protection
devices: a systematic review of the literature,” Stroke, vol. 34, no. 6, pp. 813-819,
2003.
[196] J. E. Barbato, E. Dillavou, M. B. Horowitz, T. G. Jovin, E. Kanal and S. David, “A
randomized trial of carotid artery stenting with and without cerebral protection,” J
Vasc Surg, vol. 47, no. 4, pp. 760-765, 2008.
[197] S. Macdonald, D. H. Evans, P. D. Griffiths, F. M. McKevitt, G. S. Venables and T.
J. Cleveland, “Filter-protected versus unprotected carotid artery stenting: a
randomised trial,” Cerebrovasc Dis, vol. 29, no. 3, pp. 282-289, 2010.
[198] F. M. McKevitt, M. S. Randall, T. J. Cleveland, P. A. Gaines, K. T. Tan and G. S.
Venables, “The benefits of combined anti-platelet treatment in carotid artery
stenting,” Eur J Vasc Endovasc Surg, vol. 29, no. 5, pp. 522-527, 2005.
[199] The Stenting and Aggressive Medical Management for Preventing Recurrent
Stroke in Intracranial Stenosis Trial Investigators, “Mechanisms of stroke after
intracranial angioplasty and stenting in the SAMMPRIS trial,” Neurosurgery, vol.
72, no. 5, pp. 777-795, 2013.
[200] The ICSS-MRI Substudy Investigators, “Characteristics of ischemic brain lesions
after stenting or endarterectomy for symptomatic carotid artery stenosis: results
from the International Carotid Stenting Study-Magnetic Resonance Imaging
Substudy,” Stroke, vol. 44, no. 1, pp. 80-86, 2013.
[201] G. S. Roubin, S. Iyer, A. Halkin, J. Vitek and C. Brennan, “Realizing the potential
of carotid artery stenting: proposed paradigms for patient selection and
procedural technique,” Circulation, vol. 113, no. 16, pp. 2021-2030, 2006.
190
[202] J. Y. Streifler, M. Eliasziw, A. J. Fox, O. R. Benavente, V. C. Hachinski and G. G.
Ferguson, “Angiographic detection of carotid plaque ulceration. Comparison with
surgical observations in a multicenter study. North American Symptomatic
Carotid Endarterectomy Trial,” Stroke, vol. 25, no. 6, pp. 1130-1132, 1994.
[203] L. Bonati, J. Ederle, J. Dobson, S. Engelter, R. L. Featherstone and P. Gaines,
“Length of carotid stenosis predicts peri-procedural stroke or death and
restenosis in patients randomised to endovascular treatment or endarterectomy,”
Int J Stroke, vol. 9, no. 3, pp. 297-305, 2014.
[204] J. R. Landis and G. G. Koch, “The measurement of observer agreement for
categorical data,” Biometrics, vol. 33, pp. 159-174, 1977.
[205] G. L. Faggioli, M. Ferri, A. Freyrie, M. Gargiulo, F. Fratesi and C. Rossi, “Aortic
arch anomalies are associated with increased risk of neurological events in
carotid stent procedures,” Eur J Vasc Endovasc Surg, vol. 33, no. 4, pp. 436-441,
2007.
[206] F. Burzotta, R. Nerla, G. Pirozzolo, C. Aurigemma, G. Niccoli, A. M. Leone, S.
Saffioti, F. Crea and C. Trani, “Clinical and procedural impact of aortic arch
anatomic variants in carotid stenting procedures,” Catheter Cardio Inte, p.
(online), 2015.
[207] A. I. Qureshi, N. Tariq, A. E. Hassan, G. Vazquez, H. M. Hussein and M. F. Suri,
“Predictors and timing of neurological complications following intracranial
angioplasty and/or stent placement,” Neurosurgery, vol. 68, no. 1, pp. 53-61,
2011.
[208] K. Gröschel, U. Ernemann, S. Schnaudigel, K. Wasser, T. Nägele and A.
Kastrup, “A risk score to predict ischemic lesions after protected carotid artery
stenting,” J Neurol Sci, vol. 273, no. 1, pp. 112-115, 2008.
[209] G. Faggioli, F. Monica, C. Rapezzi, C. Tonon, L. Manzoli and A. Stella,
“Atherosclerotic aortic lesions increase the risk of cerebral embolism during
carotid stenting in patients with complex arch anatomy,” J Vasc Surg, vol. 49, no.
1, pp. 80-85, 2009.
191
[210] Z. Ghogawala, M. Westerveld and S. Amin-Hanjani, “Cognitive outcomes after
carotid revascularization: the role of cerebral emboli and hypoperfusion,”
Neurosurgery, vol. 62, no. 2, pp. 385-395, 2008.
[211] A. Altinbas, M. J. van Zandvoort, E. van den Berg, L. M. Jongen, A. Algra and F.
M. Moll, “Cognition after carotid endarterectomy or stenting: a randomized
comparison,” Neurology, vol. 77, no. 11, pp. 1084-1090, 2011.
[212] A. Altinbas, M. J. van Zandvoort, E. van den Berg, A. Algra, G. J. de Borst, J.
Hendrikse, P. J. Nederkoorn, L. H. Bonati, M. M. Brown, J. Kappelle and B. H.
van der Worp, “The effect of white matter lesions on cognition after carotid
revascularization,” JNS, vol. 334, no. 1, pp. 77-82, 2013.
[213] T. Saam, M. S. Ferguson, V. L. Yarnykh, N. Takaya, D. Xu, N. L. Polisaar, T. S.
Hatsukami and C. Yuan, “Quantitative Evaluation of Carotid Plaque Composition
by In Vivo MRI,” Arteriosclerosis, Thrombosis, and Vascular Biology, vol. 25, pp.
234-239, 2005.
[214] M. Bendszus and G. Stoll, “Silent cerebral ischaemia: hiddden fingerprints of
invasive medical procedures,” Lancet Neurol, vol. 5, no. 4, pp. 364-372, 2006.
[215] M. Cosottini, M. C. Michelassi, M. Puglioli, G. Lazzarotti, G. Orlandi and G.
Parenti, “Silent cerebral ischemia detected with diffusion-weighted imaging in
patients treated with protected and unprotected carotid artery stenting,” Stroke,
vol. 36, no. 11, pp. 2389-2393, 2005.
[216] S. E. Vermeer, N. D. Prins, T. den Heijer, A. Hofman, P. J. Koudstaal and M. M.
B. Breteler, “Silent brain infarcts and the risk of dementia and cognitive decline,”
N Engl J Med, vol. 348, no. 13, pp. 1215-1222, 2003.
[217] N. Akutsu, K. Hosoda, A. Fujita and E. Kohmura, “A preliminary prediction model
with MR plaque imaging to estimate risk for new ischemic brain lesions on
diffusion-weighted imaging after endarterectomy or stenting in patients with
carotid stenosis,” Am J Neuroradiol, vol. 33, no. 8, pp. 1557-1564, 2012.
[218] N. J. Wimmer, R. W. Yeh, D. E. Cutlip and L. Mauri, “Risk prediction for adverse
events after carotid artery stenting in higher surgical risk patients,” Stroke, vol. 43,
192
no. 12, pp. 3218-3224, 2012.
[219] F. Hansen, P. Mangell, B. Sonesson and T. Länne, “Diameter and compliance in
human common carotid artery - variations with age and sex,” Ultrasound Med
Biol, vol. 21, no. 1, pp. 1-9, 1995.
[220] N. R. Cook, “Use and misuse of the receiver operating characteristic curve in risk
prediction,” Circulation, vol. 115, no. 7, pp. 928-935, 2007.
[221] G. S. Roubin, G. New, S. S. Iyer, J. J. Vitek, N. Al-Mubarak and M. W. Liu,
“Immediate and late clinical outcomes of carotid artery stenting in patients with
symptomatic and asymptomatic carotid artery stenosis: a 5-year prospective
analysis,” Circulation, vol. 103, no. 4, pp. 532-537, 2001.
[222] R. W. Hobson II, V. J. Howard, G. S. Roubin, T. G. Brott, R. D. Ferguson and J.
J. Popma, “Carotid artery stenting is associated with increased complications in
octogenarians: 30-day stroke and death rates in the CREST lead-in phase,” J
Vasc Surg, vol. 40, no. 6, pp. 1106-1111, 2004.
[223] The CAPTURE Trial Collaborators, “The CAPTURE registry: predictors of
outcomes in carotid artery stenting with embolic protection for high surgical risk
patients in the early post-approval setting,” Catheter Cardiovasc Interv, vol. 70,
no. 7, pp. 1025-1033, 2007.
[224] R. C. Lam, S. C. Lin, B. DeRubertis, R. Hynecek, K. C. Kent and P. Faries, “The
impact of increasing age on anatomic factors affecting carotid angioplasty and
stenting,” J Vasc Surg, vol. 45, no. 5, pp. 875-880, 2007.
[225] B. C. Eikelboom, T. R. Riles, R. Mintzer, F. G. Baumann, G. DeFillip and J. Lin,
“Inaccuracy of angiography in the diagnosis of carotid ulceration,” Stroke, vol. 14,
no. 6, pp. 882-885, 1983.
[226] M. Eliasziw, J. Y. Streifler, A. J. Fox, V. C. Hachinski, G. G. Ferguson and H. J.
Barnett, “Significance of plaque ulceration in symptomatic patients with high-
grade carotid stenosis. North American Symptomatic Carotid Endarterectomy
Trial,” Stroke, vol. 25, no. 2, pp. 304-308, 1994.
193
[227] F. Kennedy, J. Dobson and D. Doig, “Is surgery getting safer? An analysis of
changes in periprocedural risk of carotid endarterectomy for symptomatic carotid
stenosis over the last 30 years,” in European Stroke Conference, London, 2013.
[228] American Heart Association / American Stroke Association, “Guidelines for the
prevention of stroke in patients with stroke or transient ischemic attack,” Stroke,
vol. 42, no. 1, pp. 227-276, 2011.
[229] American Heart Association / American Stroke Association, “Guidelines for the
primary prevention of stroke,” Stroke, vol. 42, no. 2, pp. 517-584, 2011.
[230] Scottish Intercollegiate Guidelines Network, “Management of patients with stroke
of TIA: assessment, investigation, immediate management and secondary
prevention,” SIGN, 2008.
[231] National Institute for Health and Care Excellence, “Diagnosis and initial
management of acute stroke and transient ischaemic attack (TIA),” NICE, 2008.
[232] H. Sillesen, P. Amarenco, M. G. Hennerici, A. Callahan, L. B. Goldstein and J.
Zivin, “Atorvastatin reduces the risk of cardiovascular events in patients with
carotid atherosclerosis,” Stroke, vol. 39, no. 12, pp. 3297-3302, 2008.
[233] H. J. Milionis, S. Giannopoulos, M. Kosmidou, V. Panoulas, E. Manios and A. P.
Kyritsis, “Statin therapy after first stroke reduces 10-year stroke recurrence and
improves survival,” Neurology, vol. 72, no. 21, pp. 1816-1822, 2009.
[234] J. N. Redgrave, J. K. Lovett, P. J. Gallagher and P. M. Rothwell, “Histological
assessment of 526 symptomatic carotid plaques in relation to the nature and
timing of ischemic symptoms: The Oxford Plaque Study,” Circulation, vol. 113,
no. 19, pp. 2320-2328, 2006.
[235] N. Takaya, C. Yuan, B. Chu, T. Saam, H. Underhill and J. Cai, “Association
between carotid plaque characteristics and subsequent ischemic cerebrovascular
events: a prospective assessment with MRI - initial results,” Stroke, vol. 37, no. 3,
pp. 818-823, 2006.
194
[236] J. K. DeMarco, H. Ota, H. R. Underhill, D. C. Zhu, M. J. Reeves and M. J.
Potchen, “MR carotid plaque imaging and contrast-enhanced MR angiography
identifies lesions associated with recent ipsilateral thromboembolic symptoms: an
in-vivo study at 3T,” Am J Neuroradiol, vol. 31, no. 8, pp. 1395-1402, 2010.
[237] N. Singh, A. R. Moody, D. J. Gladstone, G. Leung, R. Ravikumar and J. Zhan,
“Moderate carotid artery stenosis: MR imaging-depicted intraplaque hemorrhage
predicts risk of cerebrovascular ischemic events in asymptomatic men,”
Radiology, vol. 252, no. 2, pp. 502-508, 2009.
[238] M. T. Truijman, M. E. Kooi, A. C. van Dijk, A. A. de Rotte, A. G. van der Kolk, M.
I. Liem, F. H. Schreuder, E. Boersma, W. H. Mess, R. J. van Oostenbrugge, P. J.
Koudstaal, L. J. Kappelle, P. J. Nederkoorn, A. J. Nederveen, J. Hendrikse, A. F.
van der Steen, M. J. Daemen and A. van der Lugt, “Plaque At RISK (PARISK):
prospective multicenter study to improve diagnosis of high-risk carotid plaques,”
Int J Stroke, p. ePub ahead of print, 2013.
[239] G. M. Biasi, A. Froio, E. B. Diethrich, G. Deleo, S. Galimberti, P. Mingazzini, A. N.
Nicolaides, M. Griffin, D. Raithel, D. Reid and M. G. Valsecchi, “Carotid plaque
echolucency increases the risk of stroke in carotid stenting: the imaging in carotid
angioplasty and risk of stroke (ICAROS) study,” Circulation, vol. 110, no. 6, pp.
756-762, 2004.
[240] “Carotid Stenting Trialists' Collaboration,” [Online]. Available: http://www.carotid-
trialists.com/. [Accessed 04 July 2014].
[241] Department of Clinical Neurology, Oxford University, “Model for predicting the risk
of ipsilateral ischaemic stroke in patients with recently symptomatic carotid
bifurcation stenosis,” [Online]. Available:
http://www.stroke.ox.ac.uk/model/form1.html. [Accessed 9th May 2014].
[242] “SPACE-2 Trial Protocol,” [Online]. Available: http://www.space-
2.de/downloads/SPACE-2_Protocol_120925_blank.pdf.
[243] N. Rudarakanchana, M. Dialynas and A. Halliday, “Asymptomatic Carotid Surgery
Trial-2 (ACST-2): rationale for a randomised clinical trial comparing carotid
endarterectomy with carotid artery stenting in patients with asymptomatic carotid
195
artery stenosis,” Eur J Vasc Endovasc Surg, vol. 38, no. 2, pp. 239-242, 2009.
[244] M. M. Brown, Cerebrovascular disease: pathophysiology, diagnosis and
management, Oxford: Blackwell, 1998, pp. 1931-1944.
[245] L. Coward, R. L. Featherstone and M. M. Brown, “Percutaneous transluminal
angioplasty and stenting for carotid artery stenosis,” Cochrane Library, 2001.
[246] E. B. Diethrich, M. Ndiaye and D. B. Reid, “Stenting in the carotid artery: initial
experience in 110 patients,” J Endovasc Surg, vol. 3, no. 1, pp. 42-62, 1996.
[247] J. S. Yadav, G. S. Roubin, S. Iyer, J. Vitek, P. King and W. D. Jordan, “Elective
stenting of the extracranial carotid arteries,” Circulation, vol. 95, no. 2, pp. 376-
381, 1997.
[248] P. W. Serruys, P. de Jaegere, F. Kiemeneij, C. Macaya, W. Rutsch and G.
Heyndrickx, “A comparison of balloon-expandable stent implantation in the
treatment of coronary artery disease,” New Engl J Med, vol. 331, no. 8, pp. 489-
495, 1994.
[249] D. L. Fischman, M. B. Leon, D. S. Baim, R. A. Schatz, M. P. Savage and I. Penn,
“A randomised comparison of coronary stent placement and balloon angioplasty
in the treatment of coronary artery disease,” New Engl J Med, vol. 331, no. 8, pp.
496-501, 1994.
[250] M. H. Wholey, M. Wholey, P. Bergeron, E. B. Diethrich, M. Henry and J. C.
Laborde, “Current global status of carotid artery stent placement,” Cathet
Cardiovasc Diagn, vol. 44, no. 1, pp. 1-6, 1998.
[251] M. B. Leon, D. S. Baim, J. J. Popma, P. C. Gordon, D. E. Cutlip and K. K. L. Ho,
“A clinical trial comparing three antithrombotic drug regimens after coronary
artery stenting,” New Engl J Med, vol. 339, no. 23, pp. 1665-1671, 1998.
[252] The CLASSICS Investigators, “Double-blind study of the safety of clopidogrel with
and without a loading dose in combination with aspirin compared with ticlopidine
in combination with aspirin after coronary stenting: the Clopidogrel Aspirin Stent
196
International Cooperative Study,” Circulation, vol. 102, pp. 624-629, 2000.
[253] M. M. Brown, R. L. Featherstone and L. J. Coward, “Carotid artery stenting with
and without cerebral protection,” Stroke, vol. 35, no. 11, pp. 2434-2435, 2004.
[254] H. Z. Flach, M. Ouhlous, J. M. Hendriks, M. R. van Sambeek, J. F. Veenland and
P. J. Koudstaal, “Cerebral ischemia after carotid intervention,” J Endovasc Ther,
vol. 11, no. 3, pp. 251-257, 2004.
[255] H. Poppert, O. Wolf, M. Resch, W. Theiss, T. Schmidt-Thieme and E. H. von
Graefin, “Differences in number, size and location of intracranial microembolic
lesions after surgical versus endovascular treatment without protection device of
carotid artery stenosis,” J Neurol, vol. 251, no. 10, pp. 1198-1203, 2004.
[256] S. Garcia-Sanchez, M. Millan-Torne, J. Capellades-Font, J. Muchart, P. Callejas
and N. Vila-Moriente, “Ischemic brain lesions following carotid revascularisation
procedures: a comparative study using diffusion-weighted magnetic resonance
imaging,” Rev Neurol, vol. 38, no. 11, pp. 1013-1017, 2004.
[257] K. Iihara, K. Murao, N. Sakai, N. Yamada, I. Nagata and S. Miyamoto, “Outcome
of carotid endarterectomy and stent insertion based on grading of carotid
endarterectomy risk: a 7-year prospective study,” J Neurosurg, vol. 105, no. 4,
pp. 546-554, 2006.
[258] B. Gossetti, R. Gattuso, L. Irace, E. Faccenna, S. Venosi and L. Bozzao,
“Embolism to the brain during carotid stenting and surgery,” Acta Chir Belg, vol.
107, no. 2, pp. 151-154, 2007.
[259] V. Lacroix, F. Hammer, P. Astarci, T. Duprez, C. Grandin and G. Cosnard,
“Ischemic cerebral lesions after carotid surgery and carotid stenting,” Eur J Vasc
Endovasc Surg, vol. 33, no. 4, pp. 430-435, 2007.
[260] H. G. Roh, H. S. Buyn, J. W. Ryoo, D. G. Na, W. J. Moon and B. B. Lee,
“Prospective analysis of cerebral infarction after carotid endarterectomy and
carotid artery stent placement by diffusion-weighted imaging,” Am J Neuroradiol,
vol. 26, no. 2, pp. 376-384, 2005.
197
[261] A. Kastrup, T. Nägele, K. Gröschel, F. Schmidt, E. Volger and J. Schulz,
“Incidence of new brain lesions after carotid stenting with and without cerebral
protection,” Stroke, vol. 37, no. 9, pp. 2312-2316, 2006.
[262] O. Wolf, P. Heider, M. Heinz, H. Poppert, T. Schmidt-Thieme and D. Sander,
“Frequency, clinical significance and course of cerebral ischemic events after
carotid endarterectomy evaluated by serial diffusion-weighted imaging,” Eur J
Vasc Endovasc Surg, vol. 27, no. 2, pp. 167-171, 2004.
198
199
22. Appendix I – ICSS post-treatment care questionnaire
200
23. Appendix II – ICSS cranial nerve palsy questionnaire
201
202
24. Appendix III – List of ICSS centres and investigators
Location of ICSS recruiting centres, number of patients recruited at each centre in
square brackets, and collaborators at each centre (PI) = local principal investigator
[82].
24.1.1 Australia
Austin Health, Heidelberg [46]: M Brooks, B Chambers (PI), A Chan, P Chu, D Clark,
H Dewey, G Donnan, G Fell, M Hoare, M Molan, A Roberts, N Roberts
Box Hill Hospital (Monash University), Melbourne [25]: B Beiles, C Bladin (PI), C
Clifford, G Fell, M Grigg, G New
Monash Medical Centre, Clayton [26]: R Bell, S Bower, W Chong, M Holt, A Saunder,
PG Than (PI)
Princess Alexandra Hospital, Brisbane [48]: S Gett, D Leggett, T McGahan (PI), J
Quinn, M Ray, A Wong, P Woodruff
Repatriation General Hospital, Daw Park, Adelaide [6]: R Foreman, D Schultz (PI), R
Scroop, B Stanley
Royal Melbourne Hospital, Melbourne [57]: B Allard, N Atkinson, W Cambell, S
Davies (PI), P Field, P Milne, P Mitchell, B Tress, B Yan
The Royal Hobart Hospital, Hobart [18]: A Beasley, D Dunbabin, D Stary, S Walker
(PI)
24.1.2 Belgium
Antwerp University Hospital, Antwerp [10]: P Cras, O d'Archambeau, JMH Hendriks
(PI), P Van Schil
AZ St Blasius, Dendermonde [5]: M Bosiers (PI), K Deloose, E van Buggenhout
AZ Sint Jan Brugge-Oostende, Campus Brugge, Brugges [18]: J De Letter, V
Devos, J Ghekiere, G Vanhooren (PI).
203
Cliniques Universitaires St Luc, Bruxelles [1]: P Astarci, F Hammer, V Lacroix, A
Peeters (PI), R Verhelst
Imelda Ziekenhuis, Bonheiden [3]: L DeJaegher (PI), A Peeters, J Verbist
24.1.3 Canada
CHUM Notre-Dame Hospital, Montreal [30]: J-F Blair, JL Caron, N Daneault, M-F
Giroux, F Guilbert, S Lanthier, L-H Lebrun, V Oliva, J Raymond, D Roy (PI), G Soulez,
A Weill
Foothills Medical Centre, Calgary [4]: M Hill (PI), W Hu, M Hudion, W Morrish, G
Sutherland, J Wong
24.1.4 Finland
Helsinki University Central Hospital, Helsinki [33]: Albäck A., Harno H., Ijäs P.,
Kaste M. (PI), Lepäntalo M., Mustanoja S., Paananen T., Porras M., Putaala J., Railo
M., Sairanen T., Soinne L., Vehmas A., Vikatmaa P.
24.1.5 Germany
Otto von Guericke University, Magdeburg [9]: M Goertler (PI), Z Halloul, M Skalej
24.1.6 Ireland
Beaumont Hospital, Dublin [4]: P Brennan, C Kelly, A Leahy, J Moroney (PI), J
Thornton
24.1.7 New Zealand
Auckland City Hospital, Auckland [40], PA Barber, R Bourchier, A Hill, A Holden, J
Stewart (PI)
24.1.8 Norway
Rikshospitalet University Hospital, Oslo [16]: SJ Bakke (PI), K Krohg-Sørensen , M
Skjelland, B Tennøe
204
24.1.9 Poland
Institute of Psychiatry and Neurology (2nd Department of Neurology &
Department of Neuroradiology) & Medical University of Warsaw (2nd Department
of General, Vascular and Oncological Surgery), Warsaw [20]: P Bialek, Z Biejat, W
Czepiel, A Czlonkowska (PI), A Dowzenko, J Jedrzejewska, A Kobayashi, M Lelek, J
Polanski
24.1.10 Slovenia
University Medical Centre, Ljubljana [12]: J Kirbis, Z Milosevic, B Zvan (PI)
24.1.11 Spain
Hospital Clinic, Barcelona [18]: J Blasco, A Chamorro (PI), J Macho, V Obach, V
Riambau, L San Roman
Parc Taulí Sabadell Hospital, Barcelona [33]: J Branera, D Canovas (PI), Jordi
Estela, A Gimenez Gaibar, J Perendreu
24.1.12 Sweden
Malmö University Hospital, Malmö [67]: K Björses, A Gottsater (PI), K Ivancev,T
Maetzsch, B Sonesson
Sodersjukhuset, Stockholm [55]: B Berg, M Delle, J Formgren, P Gillgren, T-B Kall, P
Konrad (PI), N Nyman, R Takolander
The Karolinska Institute, Stockholm [5]: T Andersson, J Malmstedt, M Soderman, C
Wahlgren, N Wahlgren (PI)
24.1.13 Switzerland
Centre Hospitalier Universitaire Vaudois, Lausanne [12]: S Binaghi, L Hirt, P Michel
(PI), P Ruchat
University Hospital Basel, Basel [94]: LH Bonati, ST Engelter, F Fluri, L Guerke, AL
Jacob, E Kirsch, PA Lyrer (PI), E-W Radue, P Stierli, M Wasner, S Wetzel
205
University Hospital of Geneva, Geneva [16]: C Bonvin, A Kalangos, K Lovblad, N
Murith, D Ruefenacht, R Sztajzel (PI)
24.1.14 The Netherlands
Academic Medical Centre, Amsterdam [56]: M Koelemaij, PJ Nederkoorn (PI), J
Reekers, YB Roos
Erasmus Medical Centre, Rotterdam [75]: JM Hendriks, PJ Koudstaal (PI), PMT
Pattynama, A van der Lugt, LC van Dijk, MRHM van Sambeek, H van Urk, HJM
Verhagen
The Haga Teaching Hospitals, The Hague [45]: CMA Bruijninckx, SF de Bruijn, R
Keunen, B Knippenberg, A Mosch (PI), F Treurniet, L van Dijk, H van Overhagen , J
Wever
Isala Klinieken, Zwolle [14]: FC de Beer, JSP van den Berg (PI), BAAM van Hasselt,
DJ Zeilstra
Medical Centre Haaglanden, The Hague [3]: J Boiten (PI), JCA de Mol van Otterloo,
AC de Vries, GJ Lycklama a Nijeholt, BFW van der Kallen
UMC St Radboud, Nijmegen [13]: JD Blankensteijn, FE De Leeuw, LJ Schultze Kool
(PI), JA van der Vliet
University Medical Centre, Utrecht [270]: GJ de Borst, GAP de Kort, LJ Kapelle (PI),
TH Lo, WPThM Mali, F Moll, HB van der Worp, H Verhagen
24.1.15 United Kingdom
Addenbrookes Hospital, Cambridge [5]: N Higgins, PJ Kirkpatrick, P Martin (PI), K
Varty
Birmingham Heartlands Hospital, Birmingham [11]: D Adam, J Bell, AW Bradbury, P
Crowe, M Gannon, MJ Henderson, D Sandler, RA Shinton (PI), JM Scriven, T Wilmink
Lancashire Teaching Hospitals NHS Trust, Preston [2]: S D'Souza, A Egun, R Guta,
S Punekar, DM Seriki (PI), G Thomson
206
Liverpool Royal Infirmary [21] and The Walton Centre, Liverpool [7]: JA Brennan,
TP Enevoldson, G Gilling-Smith (PI), DA Gould, PL Harris, RG McWilliams, H-C
Nasser, R White
Manchester Royal Infirmary, Manchester [2]: KG Prakash, F Serracino-Inglott, G
Subramanian (PI), JV Symth, MG Walker
Newcastle Acute Hospitals NHS Foundation Trust, Newcastle-Upon-Tyne [108]: M
Clarke, M Davis, SA Dixit, P Dorman (PI), A Dyker, G Ford, A Golkar, R Jackson, V
Jayakrishnan, D Lambert, T Lees, S Louw, S Macdonald, AD Mendelow, H Rodgers, J
Rose, G Stansby, M Wyatt
North Bristol NHS Trust, Frenchay Hospital, Bristol [13]: T Baker, N Baldwin (PI), L
Jones, D Mitchell, E Munro, M Thornton
Royal Free Hospital, London [1]: D Baker, N Davis, G Hamilton (PI), D McCabe, A
Platts, J Tibballs
Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield [151]: J Beard, T
Cleveland, D Dodd, P Gaines, R Lonsdale, R Nair, A Nassef, S Nawaz, G Venables
(PI)
St George’s University of London and St George’s NHS Healthcare Trust, London
[58]: A Belli, A Clifton, G Cloud, A Halliday, H Markus (PI), R McFarland, R Morgan, A
Pereira, A Thompson
St Mary’s Hospital, Imperial College Healthcare NHS Trust, London [13]: J
Chataway (PI), N Cheshire, R Gibbs, M Hammady, M Jenkins, I Malik, J Wolfe
University College London Hospitals NHS FoundationTrust, London [51]: M
Adiseshiah, C Bishop, S Brew, J Brookes, MM Brown (PI), R Jäger, N Kitchen
University Hospital of South Manchester, Wythenshawe, Manchester [58]: R
Ashleigh, S Butterfield, GE Gamble, C McCollum (PI), A Nasim, P O'Neill, J Wong
Western Infirmary, Glasgow [5]: RD Edwards, KR Lees, AJ MacKay, J Moss (PI), P
Rogers
207
25. Appendix IV – ICSS trial protocol
Protocol summary 25.1
25.1.1 Background
Clinical trials have shown that carotid surgery prevents stroke but also has significant
morbidity. Stenting has become an established alternative treatment for coronary and
peripheral vascular disease and has the advantage of avoiding general anaesthesia
and neck incision. In July 1997, randomization was completed in the Carotid and
Vertebral Artery Transluminal Angioplasty Study (CAVATAS). The results did not show
a difference in the major risks or benefits of carotid angioplasty and surgery, but the
trial did show that both methods still carry a significant risk of causing a stroke.
Techniques of carotid angioplasty have improved and stenting is increasingly used.
The International Carotid Stenting Study (ICSS or CAVATAS 2) is a follow-on study to
CAVATAS designed as an international, multicentre, randomized trial, which will
evaluate stenting of carotid artery stenosis in patients with cerebrovascular disease.
25.1.2 Centre requirements
A neurologist or physician with an interest in stoke, a surgeon with expertise in carotid
endarterectomy and an interventionalist with expertise in carotid angiography and the
techniques of angioplasty and stenting.
25.1.3 Inclusion criteria
Symptomatic atheromatous carotid stenosis, >50% by NASCET criteria, suitable for
stenting and surgical endarterectomy.
25.1.4 Treatments
Patients will be randomised in equal proportions to be treated by carotid
endarterectomy or stenting. New designs of stents, filters and protection devices will be
incorporated into the study to allow tracking of new technology if approved by the
Steering Committee. Surgery can be performed with local or general anaesthesia.
208
25.1.5 Sample size
N=1500 patients from fully enrolled centres. Sample size calculations show that the
95% confidence intervals will be 3.0 percentage points for the outcome measure of 30
day stroke, myocardial infarction and death rate and 3.3 percentage points for the
outcome measure of death or disabling stroke during follow-up.
25.1.6 Primary outcome measure
Long term survival free of disabling stroke
25.1.7 Secondary outcome measures
Any stroke, myocardial infarction or death within 30 days of treatment,
treatment-related cranial nerve palsy or haematoma
Stenosis (>70%) and occlusion on ultrasound follow-up
Transient ischaemic attack
Stroke during follow-up
Further treatment procedure
Quality of life and economic measures
Background 25.2
Stroke is the major cause of acquired adult physical disability and is responsible for
12% of all deaths in the UK. Reducing the burden of stroke is one of the priorities of the
recent government white paper, Saving Lives: Our Healthier Nation. In Europe alone,
there are approximately one million new cases of stroke a year. Atherosclerotic
stenosis of the carotid artery is an important cause of stroke, which may be heralded by
a transient ischaemic attack (TIA) or minor stroke, which recovers without serious
disability. The risk of recurrent stroke in recently symptomatic patients with severe
carotid stenosis is as high as 28% over two years. The European Carotid Surgery Trial
(ECST) and the North American Symptomatic Carotid Endarterectomy Trial (NASCET)
have demonstrated convincingly that this risk is reduced significantly by carotid
endarterectomy [36] [35]. Carotid surgery has therefore become a standard treatment
for these patients. However, the trials showed a significant risk of stroke or death
resulting from surgery of between 6 and 8%. Surgery also caused significant morbidity
from myocardial infarction during the general anaesthetic used in most centres, and
minor morbidity, including cranial nerve palsy and wound haematoma, from the
209
incision. An increasing number of surgeons are performing carotid endarterectomy
under local anaesthesia in the belief that it reduces the risks, although there is currently
little evidence to support this practice, until the data from the General Anaesthesia
versus Local Anaesthesia for Carotid Endarterectomy (GALA) trial are reported.
Stenting is a new method of treating carotid stenosis which has evolved from the
technique of percutaneous transluminal angioplasty (PTA). Stenting avoids some of the
hazards of surgery and has become an established treatment for peripheral and
coronary artery stenosis. Stenting is less invasive than carotid endarterectomy and has
advantages in terms of patient comfort, because the procedure avoids an incision in
the neck, and is usually conducted under local anaesthesia. Hospital stay need only be
for 24 hours after treatment if uncomplicated. When given the choice, stenting is
preferred by many patients. On the other hand, stenting does not remove
atheromatous plaque, has not been shown to prevent stroke and may have an
unacceptable incidence of restenosis. We therefore propose a multicentre randomised
trial to compare carotid stenting with carotid surgery.
25.2.1 Previous work in the field
Percutaneous transluminal angioplasty
A number of groups have published series of patients with carotid stenosis treated by
PTA. The cumulative total of patients in these series is over 1000, with a reported
major complication rate of less than 5% at the time of the procedure [244]. These data
suggested that carotid PTA has a similar risk to carotid surgery, but the results could
not be taken as definitive because none of the data were from randomised trials. We
therefore started a randomised trial, known as the Carotid and Vertebral Artery
Transluminal Angioplasty Study (CAVATAS) in l992. We completed randomisation in
July 1997.
Results of CAVATAS
560 patients were entered, from 24 centres in the UK, Australia, Canada, Finland,
Germany, Italy, Spain, Switzerland, and the USA. Patients with carotid stenosis
suitable for surgery were randomised between PTA (n=251) and carotid surgery
(n=253). Patients with carotid stenosis unsuitable for surgery (n=40) and patients with
vertebral artery stenosis (n=16) were separately randomised between PTA and
medical care alone. The number of patients in these last 2 groups was too small to
form any firm conclusions. The analysis has therefore been restricted to the 504
210
patients with carotid stenosis randomised between PTA and surgery [75]. Baseline
variables were well matched. Almost all patients had severe stenosis (mean 86%). The
30-day outcome events were almost identical in the two groups with a rate of death or
any stroke lasting more than 7 days of 10.0% after angioplasty and 9.9% after surgery,
giving a hazard ratio of 1.01 (95% CI 0.56 to 1.81, p=NS). Analysis of the other risks of
treatment has confirmed that PTA was safer than surgery in terms of minor morbidity.
Cranial or peripheral nerve palsy was reported in 9% of surgical patients, but not in any
PTA patients (p<0.01). Haematoma requiring operation or prolonging hospital stay was
reported in 7% of surgical patients compared with 1% of PTA patients (p<0.01). PTA
also appeared safer than surgery with regard to perioperative myocardial infarction,
which occurred in 0.8% of surgical patients, but not in any PTA patients. Survival
analysis from randomisation showed no difference in outcome events of ipsilateral
stroke and any disabling stroke or death during follow up for up to 3 years with very few
events in either arm after the treatment period, suggesting that both treatments were
equally effective at preventing stroke. However, 19% of PTA patients had stenosis of
>70% or occlusion by ultrasound criteria at 12 months after randomisation compared
with 5% of surgical patients (p<0.01). Restenosis was not associated with new
symptoms, but long-term follow up is limited.
Causes and timing of stroke in CAVATAS
The cause of stroke within 30 days of first treatment in CAVATAS was cerebral
infarction in 22 patients in the PTA group and 20 patients in the surgery group. Primary
cerebral haemorrhage caused the other three strokes in the PTA group and two
strokes in the surgery group. All but one stroke was ipsilateral to the randomised artery.
Surprisingly, a significant proportion of these treatment-related strokes were delayed
after the day of treatment. Eight (36%) of the strokes in PTA patients occurred between
the second and 21st day after treatment. Delayed stroke was also found in six (27%) of
the surgical group between the third and 10th day after operation. Delayed stroke may
account for the relatively high rate of 30-day morbidity in CAVATAS at 10% compared
to 7.5% in ECST and 5.8% in NASCET.
Carotid stenting
Stents suitable for carotid use have only become available recently. The CAVATAS
Steering Committee decided to allow the use of stents at the discretion of the
interventionist. Stents were used in 55 patients randomised to PTA, usually as a
secondary procedure i.e. after initial balloon dilation. The indication for using a stent in
211
these cases was usually an inadequate angiographic result and in some cases stents
were deployed because of stroke at the time of full balloon inflation, as a ‘bail-out’
procedure. Only one stroke occurred at the time of stent deployment (1.8%), although
there were a small number of delayed strokes after stenting.
The need for a trial of carotid stenting
It would be inappropriate to use the results of CAVATAS to propose the widespread
introduction of PTA for the treatment of carotid stenosis as an alternative to surgery,
because the 95% confidence interval surrounding the 10% risk of any stroke within 30
days of treatment in the surgical and angioplasty groups is ±5%. Nevertheless, the
results support the need for further randomised studies. The interventional technique
used to treat carotid stenosis has evolved over the 7 years since we started CAVATAS,
from the use of simple inflatable balloon catheters at the beginning of the trial to the
increasing use of stenting towards the end of the trial. Initially stents were used only as
a secondary procedure after full balloon inflation for inadequate results or
complications of treatment. The desire to prevent these complications and superior
early results in stented patients has led to the increasing use of the technique of
primary stenting in which the intention is to deploy a stent in every patient before
dilation (but after pre-dilatation to allow the atraumatic passage of the stent) of the
artery [245]. Primary stenting is now accepted as best practice and has become the
radiological technique of choice for carotid stenosis, replacing balloon angioplasty.
Advantages of carotid stenting
The majority of major strokes after carotid PTA are the result of dissection of the
carotid artery at the time of balloon inflation with subsequent thrombosis. It is believed
that stenting is safer than simple balloon angioplasty because embolisation, dissection
and closure of the carotid artery are less likely to occur [246] [247]. The subgroup
analysis of stented patients in CAVATAS is consistent with this suggestion. The
adverse consequences of dissection are minimised, because the stent maintains
laminar flow across the stenosis and seals the site of dissection, preventing a free
intimal flap. In addition, the stent mesh limits the size of any thrombus or atheromatous
debris that may be dislodged from the plaque at the time of dilation of the artery.
Superior dilation achieved by stenting compared with balloon angioplasty may also
reduce the rate of stroke in the early post-treatment period. In the coronary circulation,
stenting has been shown to produce superior outcomes compared with balloon
angioplasty [248] [249]. Individual case series suggest that carotid stenting has a
212
similar rate of procedural stroke to that of carotid surgery [246] [247], while a recent
registry reported a total of 2048 patients from 24 centres undergoing carotid stenting
with a complication rate of stroke and death within 30 days of treatment of 5.8% [250].
Disadvantages of carotid stenting
Although acceptable safety at the time of stenting has been suggested by the case
series and registry data, stenting has not been subjected to a randomised trial in
comparison to conventional surgical treatment and has not been demonstrated to
prevent stroke, which is the aim of treatment. Stenting does not remove atheromatous
plaque and stents may stimulate neo-intimal hyperplasia. In the long term it is likely that
the rate of restenosis will be greater after stenting than after carotid surgery, which
could well result in an unacceptable rate of long-term stroke recurrence. There is an
important need to establish the efficacy of carotid stenting in comparison to surgery
before the technique is widely introduced without adequate trial based evidence.
Antiplatelet therapy
In cardiological practice, ischaemic complications during coronary stenting have been
shown to be significantly reduced by using a combination of two antiplatelet agents,
ticlopidine and aspirin. In one coronary trial, stent thrombosis was reduced from 3.6%
in patients assigned aspirin alone down to 0.5% in patients assigned aspirin and
ticlopidine [251]. A recently completed trial has shown that similar results with less risk
of side effects can be achieved during coronary stenting by using the combination of
clopidogrel with aspirin [252]. It is likely that this combination would also reduce the
risks of stroke during carotid stenting. A pilot study is currently being carried out at one
of the centres to establish the safety of the combination of clopidogrel and aspirin given
before and for 30 days after carotid stenting. It is likely that this will become standard
therapy. Most surgeons currently believe that combination antiplatelet therapy during
surgery is hazardous because of excess bleeding.
Economic and quality of life considerations
Quality of life and general health status were assessed in CAVATAS using the SF36
and EuroQol EQ-5D questionnaires. These showed a similar quality of life for patients
randomised to either treatment. Operating and radiology suite costs were similar in a
sample of patients at two UK centres, but surgery was associated with a longer hospital
stay and greater use of ITU beds. Surgery was therefore considerably more expensive
than angioplasty (mean difference £946). However, the mean cost of an angioplasty
213
increased from £1086 to £1864 if a stent was used. The use of stents in every case is
therefore likely to increase the costs of stenting close to that of surgery, but this might
be counterbalanced by performing carotid stenting as a day case procedure. Surgical
length of stay is also declining. Follow up costs might be very different if restenosis is
more frequent in one arm. Economic analysis will therefore be an important component
of ICSS.
25.2.2 Aims of ICSS
To compare the risks, benefits and cost effectiveness of a treatment policy of referral
for carotid stenting compared with referral for carotid surgery.
Trial design 25.3
ICSS is an international, multicentre, randomised, controlled, open, prospective clinical
trial comparing carotid surgery with carotid stenting.
25.3.1 Participating centre requirements
Each centre must have a neurologist or physician with an interest in stroke who will see
patients prior to randomisation and for follow up. Carotid endarterectomy must be
carried out by designated surgeons with expertise in the operation. Carotid stenting will
be carried out by designated consultant interventionists with expertise in carotid
angiography and the techniques of angioplasty and stenting. Good collaboration
between the neurologists, surgeons and interventionists is essential and centres should
have regular neurovascular meetings. Attendance at training sessions in carotid
stenting provided by credentialing centres will be required for all interventionists prior to
participation. Participating centres will be required to submit curriculum vitae for all
participating clinicians and an audit of recent carotid surgery and PTA / stenting results.
An accreditation committee will decide if they have appropriate experience and
expertise to join the study. As a guide, surgeons and interventionists will be expected
to show a stroke and death rate within 30 days of treatment, consistent with the centres
in ECST who had an average rate of 7.0% with a 95% confidence interval of 5.8 to
8.3% [36]. Surgeons will be expected to have performed a minimum of 50 carotid
operations with a minimum annual rate of at least 10 cases per year. Interventionists
will similarly be expected to have performed a minimum of 50 stenting procedures, of
which at least 10 should be in the carotid territory. Centres where there is little or no
experience of carotid stenting may join ICSS for a probationary period in order to gain
the minimum experience of ten carotid stenting procedures required to join the trial
214
fully. The results in patients randomised during the probationary period will be analysed
separately.
All centres will have to provide proof of ethical committee approval for the study before
commencing randomisation.
25.3.2 Probationary centres
Probationary centres will be required to fulfil all the other requirements for entry, but will
not have to provide audited data on ten carotid stenting procedures initially.
Probationary centres will randomise patients within the ICSS protocol between surgery
and stenting. Individual interventionists who are not able to satisfy the credentialing
requirements will be identified as probationary investigators. Stenting procedures
carried out during the probationary period must be proctored by an experienced carotid
interventionist, until the proctor is satisfied that the interventionist(s) at the centre can
satisfactorily carry out procedures unproctored. Probationary interventionists will
become fully enrolled in ICSS when both the proctor is satisfied that the interventionist
can perform procedures unsupervised and the interventionist has 10 or more
successfully completed cases in the trial, with an acceptable complication rate. When
an investigator has done sufficient successful procedures, the trial office will get
comments from the relevant proctor, and then have any decision to promote the
investigator or centre signed off by the chair of steering committee.
25.3.3 Proctoring
Proctors for probationary centres will be approved by the accreditation committee in
consultation with the probationary centre via the central ICSS office. Probationary
centres may suggest an appropriate proctor, but he or she will require prior approval
from the accreditation committee, based on review of the proctor’s experience of
carotid stenting. It is the responsibility of the probationary interventionist to make
contact with an approved ICSS proctor and to ensure a convenient date is organised
for the stenting procedure at which the proctor can be present. Copies of the relevant
radiology should be available for the proctor for review prior to starting the stenting
procedure. This should be done prior to randomisation if there was any doubt about the
suitability of the patient for stenting. In the event of a centre requiring proctoring for
surgery the same procedure will apply.
It is the responsibility of the probationary interventionist and the proctor in discussion to
ensure the lesion is appropriate for treatment (e.g. sufficiently severe), that the patient
215
has received appropriate premedication (e.g. a combination of clopidogrel and aspirin)
and that the lesion is suitable for stenting. They should agree the type, range and sizes
of equipment required and the probationary interventionist should ensure that this
equipment is available to complete the procedure. If any of these conditions are not
met, the procedure should be abandoned and if appropriate rescheduled for another
occasion.
Catheter or arch angiography is not required in ICSS prior to randomisation if the
centre does not routinely perform angiography prior to treatment. However, the centre
interventionist, the proctor and the patient should be aware that if preliminary
angiography at the time of planned stenting shows a lesion which is not suitable for
stenting, the procedure should be abandoned and the patient referred for surgery, or
continued medical management. This type of cross over is envisaged in the trial
design.
Where a centre has an adequately qualified surgeon and interventionist they may
supervise surgeons and interventionists at the same centre whose experience would
not initially qualify them for the trial until they have gained sufficient experience. These
new investigators must enrol with the central ICSS office (see participating centre
requirements).
25.3.4 Inclusion criteria
Symptomatic extracranial, internal or bifurcation atheromatous carotid artery
stenosis that is suitable for both stenting and surgery and is deemed by the
randomising clinician to require treatment
The severity of the stenosis of the randomised artery should be at least 50% (as
measured by NASCET method or non-invasive equivalent)
Symptoms must have occurred in the 12 months before randomisation. It is
recommended that the time between symptoms and randomisation should be
less than 6 months, but patients with symptoms occurring between 6 and 12
months may be included if the randomising physician considers treatment
indicated
The patient must be clinically stable following their most recent symptoms
attributable to the stenotic vessel
Patients must be willing to have either treatment, be able to provide informed
consent, and be willing to participate in follow up
216
Patients must be able to undergo their allocated treatment as soon as possible
after randomisation
Any age greater than 40 may be included. There is no upper age limit
Patients should only be randomised if the investigator is uncertain which of the
two treatments is best for that patient at that time
25.3.5 Exclusion criteria
Patients refusing either treatment
Patients unable or unwilling to give informed consent
Patients unwilling or unable to participate in follow up for whatever reason
Patients who have had a major stroke with no useful recovery of function within
the territory of the treatable artery
Patients with a stenosis that is known to be unsuitable for stenting prior to
randomisation because of one or more of:
o Tortuous anatomy proximal or distal to the stenosis
o Presence of visible thrombus
o Proximal common carotid artery stenotic disease
o Pseudoocclusion (‘string sign’)
o Patients not suitable for surgery due to anatomical factors e.g. high
stenosis, rigid neck
Patients in whom it is planned to carry out coronary artery bypass grafting or
other major surgery within 1 month of carotid stenting or endarterectomy
Carotid stenosis caused by non-atherosclerotic disease e.g. dissection,
fibromuscular disease or neck radiotherapy.
Previous carotid endarterectomy or stenting in the randomised artery
Patients in whom common carotid artery surgery is planned
Patients medically not fit for surgery
Patients who have a life expectancy of less than two years due to a pre-existing
condition, e.g. cancer
25.3.6 Non-randomised patients
An anonymised log will be kept of patients undergoing treatment for carotid stenosis by
the trial investigators but not randomised at the participating centres. Patients
undergoing stenting but not randomised should also be included on a suitable registry,
such as EUROCAST.
217
25.3.7 Consent
Written witnessed, informed consent will be obtained from all patients and a copy must
be retained by the randomising centre. All patients will be provided with a written
explanation of the study.
25.3.8 Randomisation
Randomisation will be by a telephone call or fax to a computerised service provided by
the Oxford Clinical Trials Service Unit. Randomisation will be stratified by centre with
minimisation of the main risk factors and balanced between the arms. Patients who
need treatment of both carotid arteries will only be randomised for the carotid artery to
be treated first. Patients can only be randomised once.
25.3.9 Investigations before randomisation
The following investigations are required: Routine haematology (FBC, platelets), blood
biochemistry (renal function, blood sugar, cholesterol), chest x-ray, ECG, brain CT or
MRI scans. The brain scan is required to exclude other pathology, to identify existing
infarcts and to provide a baseline reference against which any subsequent infarction or
haemorrhage can be assessed. Copies of the CT or MRI scans should be sent to the
ICSS office.
25.3.10 Carotid imaging
Mandatory investigation is required for entry into the study to confirm the presence and
severity of the ipsilateral stenosis and to assess contralateral carotid disease. The
following are acceptable:
Arch arteriogram showing both carotid bifurcations
Selective catheter carotid angiography showing the randomised carotid artery
with non-invasive investigation of the contralateral carotid bifurcation
Bilateral magnetic resonance carotid angiograms together with a concordant
ultrasound scan
Bilateral spiral CT angiograms together with a concordant ultrasound scan
Bilateral duplex and Doppler ultrasound scan, only if this is standard practice to
treat on the basis of ultrasound alone in individual centres and the centre has
218
been able to provide proof of the reliability of their ultrasonographic imaging
through clinical audit
The following data from the pre-randomisation imaging will be sent to the Central Office
for review:
A copy of the written reports of the studies
A film copy of the view of the vessel to be treated showing the stenosis at its
most severe
A film copy of the view of the contralateral vessel showing any stenosis at its
most severe
Velocity data from the ultrasound examination
Patients who are randomised to stenting after ultrasound or other non-invasive
investigation, in which subsequent angiography, prior to stenting, reveals one or more
exclusion criteria should be treated by surgery, if appropriate, or medical care only if
surgery is not appropriate (e.g. because the stenosis is less than 50%). These patients
will continue follow up in the trial and will be analysed on an intention to treat basis. A
similar approach should be taken to patients randomised to surgery in whom
contraindications to surgery emerge after randomisation.
25.3.11 Ultrasound
Ultrasound study of the carotid artery to be treated will be performed at or before
randomisation, at one month after treatment and then annually after randomisation in
all patients. The following information is required for each study: Peak systolic velocity
of internal carotid artery (PSV ICA), end diastolic velocity of internal carotid artery (EDV
ICA), peak systolic velocity of common carotid artery (PSV CCA). The accuracy of
individual ultrasound laboratories will be audited by comparing the pretreatment
ultrasound examination against catheter angiography films, which will be available in
patients randomised after angiography and in all the patients treated by stenting.
25.3.12 Baseline data
Baseline data collected at randomisation will include demographic data; existing
medical risk factors; neurological symptoms including an assessment of disability using
219
the Modified Rankin Scale; current antiplatelet therapy and blood pressure. Films
and/or reports of pre-randomisation imaging as detailed above and in all cases the
results of Doppler ultrasound as detailed below are required to allow assessment of
any subsequent stenosis.
25.3.13 Baseline assessment
Patients will be seen by the study neurologist or physician interested in stroke prior to
randomisation to confirm suitability for the study.
25.3.14 Stenting protocol
Stenting will be carried out as soon as possible after randomisation using percutaneous
transluminal interventional techniques from the femoral, brachial or common carotid
artery by a designated interventional consultant using an appropriate stent. A cerebral
protection system should be used whenever the operator thinks one can be safely
deployed. Stents and other devices used in the trial must be CE marked and approved
by the Steering Committee. Pre-medication will be discretionary. The combination of
aspirin and clopidogrel is recommended as the antiplatelet regime of choice to cover
the period of stenting and for a minimum of 4 weeks afterwards. Intra-procedural
heparin is mandatory at a dose determined by the operator, post-procedural heparin
may be given according to clinical requirements. Patients should be monitored for
changes in their neurological status and heart rate throughout the procedure. If femoral
or brachial access is being used a long sheath introducer or a guiding catheter is
placed in the common carotid artery allowing pre-dilation and stent placement under
direct arteriographic imaging. Atropine, or a similar agent, must be administered prior to
stent deployment to counteract any effects on the carotid artery baroreceptors, which
could lead to severe bradycardia and/or asystole. Virtually all patients will require pre-
dilatation of the stenosis by balloon angioplasty prior to stent deployment. This will
minimise the embolic load caused by passage of the endoluminal stent through the
stenosis. The size of the pre-dilatation balloon will be determined by the size of the
delivery system being used. Further balloon dilation of the stent will usually be required
to ensure apposition of the stent against the arterial wall. Angiographic images
showing the stenosis at its most severe prior to stenting and the same view and any
other view that demonstrates the maximum residual stenosis after stenting must be
sent to the Central Office. Details of the procedure, including all peri-procedural
complications, drug therapy and devices used in the procedure, must be reported and
220
the stenting and cerebral protection technical data sheet returned to the trial Central
Office.
25.3.15 Endarterectomy protocol
Endarterectomy is to be done as soon as possible after randomisation by a designated
consultant surgeon who has been approved by the Credentials Committee. It is to be
carried out using whichever procedures are standard at the individual centre, including
the use of local or general anaesthesia, shunts or patches as required by the operating
surgeon. Standard or eversion endarterectomy may be performed.
25.3.16 Reporting of suspected problems with surgical or stenting
techniques at individual centres
If the local investigator or other member of the team at a trial centre has concern about
the outcome of their trial procedures they should inform the ICSS trial office which will
organise a blinded assessment of the relevant outcome events. This will be submitted
by the central office to the chairman of the data monitoring committee who may
recommend further action, such as suspending randomisation at the centre. Similarly,
the database manager at the trial office will monitor outcome events and if there are
two consecutive deaths or three consecutive major events at a single centre within 30
days of treatment in the same arm of the study, then assessment of the events will be
triggered. A cumulative major event or death rate of more than 10% over 20 cases
would also trigger careful assessment of the relevant outcome events.
25.3.17 Medical treatment
All patients will receive best medical care including antiplatelet therapy or
anticoagulation (when appropriate) and control of medical risk factors such as
hypertension, smoking and hyperlipidaemia before treatment and throughout the period
of follow up.
25.3.18 Prevention of thrombosis
Therapy to prevent thrombosis during or soon after surgery or stenting will be
prescribed according to standard practice in each centre. This may include heparin,
dextran, aspirin, dipyridamole, ticlopidine, clopidogrel, or a combination of aspirin and
another antiplatelet agent. Glycoprotein IIb/IIIa antiplatelet receptor antagonists will not
be used routinely.
221
25.3.19 Follow-up
Patients will be followed up by a neurologist or a physician interested in stroke at the
participating centres at 30 days after treatment, 6 months after randomisation and then
annually after randomisation. All post-procedural complications occurring within thirty
days after the procedure will be reported to the central office at the 30 day follow up. At
each visit, levels of stroke-related disability will be assessed using the modified Rankin
Scale and any relevant outcome events will be notified to the Central Office. A Doppler
ultrasound will be used to measure carotid arterial diameter to assess patency at one
month after treatment and then annually after randomisation. In addition, ultrasound re-
examination and CT or MRI scan should be performed in patients who have any
transient ischaemic events and/or stroke during follow up. The duration of follow up will
be a minimum of 5 years (or until termination of the trial if earlier). At the 5 year follow
up, patients will be asked if they are willing to continue follow up, in which case annual
follow up will continue up to a maximum of 10 years from randomisation.
25.3.20 Sample size calculations and recruitment
The planned sample size is 1500. We do not anticipate any large difference in the
principal outcome between surgery and stenting. We propose to estimate this
difference and present a confidence interval for difference in 30-day death, stroke or
myocardial infarction and for three-year survival free of disabling stroke or death. For
1500 patients, the 95% confidence interval will be the observed difference ±3.0% for
the outcome measure of 30 day stroke, myocardial infarction and death rate and ±3.3%
for the outcome measure of death or disabling stroke over three years follow up.
However, the trial will have the power to detect major differences in the risks of the two
procedures, for example if stenting proves to be much riskier than surgery or
associated with more symptomatic restenosis. The differences detectable with power
80% are 4.7% for 30 day outcome and 5.1% for survival free of disabling stroke.
Similar differences are detectable for secondary outcomes. We expect to achieve this
recruitment within 6 years.
222
Principal research questions to be addressed 25.4
25.4.1 Primary analysis
What is the difference in the long-term rate of fatal or disabling stroke in any
territory of patients with severe symptomatic stenosis after randomisation to a
policy of carotid stenting compared to surgery?
25.4.2 Secondary analysis
What are the differences in mortality and morbidity within 30 days of carotid
stenting compared to surgery?
What is the rate of symptomatic and asymptomatic restenosis after carotid
stenting compared to surgery?
What are the differences in the rate of ipsilateral stroke during follow-up after
carotid stenting compared to surgery?
What is the cost-effectiveness of carotid stenting compared to surgery?
What are the risk factors for stroke within 30 days and during long term follow
up (including those related to age, gender, symptoms, imaging, centre and
technique)?
25.4.3 Outcome events
Any stroke or death
Transient ischaemic attack
Myocardial infarction within 30 days of treatment
Cranial nerve palsy within 30 days of treatment
Haematoma caused by treatment requiring surgery or transfusion or prolonging
hospital stay
Stenosis greater than 70% or occlusion during follow-up
Further treatment of the randomised artery by interventional radiology
techniques or surgery after the initial attempt
Quality of life, health status and health service costs (see paragraph below)
25.4.4 Outcome event reporting
Outcome events will be documented in detail by the investigating centre, censored
after receipt at the central office to remove clues as to the treatment received, and then
223
adjudicated by an independent neurologist. Patients suffering stroke should have a CT
or MRI brain scan as soon as possible after the event. A film copy of this, together with
a film copy of the pre-randomisation scan (if done) should be submitted together with a
report of the event. The event report should include copies of discharge summaries,
death certificates and post mortem results if relevant. Deaths of UK patients will be
tracked by flagging patients against the UK Registry of Births and Deaths. Disability
after stroke and cranial nerve palsy will be assessed 30 days and six months after
treatment or onset, using the modified Rankin Scale. Duration of symptoms will be
recorded and outcome events will be classified as disabling if the Rankin score is 3 or
more at six months.
Learning curve 25.5
Carotid stenting is a new procedure, while the techniques of carotid surgery are well
established. It is likely that there will be a learning curve for carotid stenting and the
results may improve with experience during the trial. However, we believe it is better
that carotid stenting should be performed as part of a randomised clinical trial at this
stage of its development, because this will ensure careful assessment and follow up of
all patients treated in the trial and supervision from the Data Monitoring and Ethics
Committee ensures that continuing treatment with the new technique remains ethical.
The influence of the early part of the learning curve for carotid stenting will be limited by
careful training of individual interventionists. The total experience of carotid PTA and
stenting of individual interventionists will be recorded prior to entry into the trial. This
will allow the average duration of the learning curve to be analysed, taking into account
the current experience of the individual interventionists. This information may have
implications for interpretation of the results of the trial and for the future training and
supervision of the procedure. Similarly, there may be improvements in individual
surgical or anaesthetic techniques during the trial.
Effect of changes in technology during the course of the study 25.6
The field of carotid stenting is an area of fast changing technology. The protocol does
not at present specify the type or manufacturer of the stents or protection devices to be
used, but devices to be used in the trial will be CE marked and approved by the
Steering Committee who will expect a peer reviewed report of device safety. More than
one device may be recommended to allow the interventionist to tailor the choice of
stent to the individual stenosis and to use new designs of stent or protection devices if
appropriate. The protocol will not specify the technique to be used during carotid
224
surgery. Decisions about the use of shunts or specific suture materials will be left to the
individual surgeon. Local or general anaesthesia will be allowed in both arms.
Technical details of surgical and stenting technique, including the manufacturer and
type of stent used, the use of local or general anaesthesia, and the use of
antithrombotic agents, will be recorded. The analysis will include a subgroup
comparison of different techniques in both arms and the data will be presented to the
DMC meetings to ensure that no one technique is significantly inferior to another.
Randomisation will use a computer program to minimise variation between centres and
over time, so that equal numbers of patients will be entered into the stenting and
surgery arms before and after any change in practice.
Health service research issues 25.7
If the trial confirms the hypothesis that carotid stenting and surgery are equivalent in
terms of the major risks of stroke and death, then the choice between the two
procedures will be determined primarily by differences between the two procedures in
other outcomes e.g. the disadvantage of a scar or cranial nerve palsy, or the effects of
surgery on health related quality of life. If these differences are minor, the choice
between the procedures will be made primarily on economic grounds. The effects of
cranial nerve palsy may be detected by a minor increase in the disability score, but it is
not easy to assess the effect of these outcomes during follow-up on clinical
examination alone. Quality of life and health status will therefore be assessed using the
EuroQol (EQ5D) questionnaire to compare patients’ feeling of well-being, health and
quality of life before and after stenting or surgery at one month, six months and annual
follow-up. The results will be analysed blind to treatment arm. The first questionnaire
will be completed at the time of randomisation and subsequent questionnaires at each
follow-up visit. The investigator performing randomisation or follow-up should ensure
the patient completes the EQ5D at the same time. The English language version of the
EQ5D has been modified to record the date on which it is completed and the patient’s
trial number. Those centres using versions in other languages should also record the
date and trial number on each completed form. If patients are too disabled to complete
the questionnaires themselves, the patient’s carer may complete them. The EQ5D
should be returned to the central office with the other trial forms.
Information on hospital resource use during the treatment and follow up, including the
type and manufacturer of the devices employed in carotid stenting procedures, will be
collected to measure treatment costs and estimate the costs of stroke and any
consequences of restenosis (e.g. re-treatment) during follow up. Unit costs will be
225
obtained from a sample of representative centres. The costs of stroke caused by
treatment are a major component of the total cost of treatment, and therefore have a
major influence on cost effectiveness. As the additional length of stay in hospital
resulting from stroke largely drives these costs, the prospective collection of length of
stay data will be designed to capture the stroke-related data in addition to direct
operative stay. The economic evaluation will address cost-effectiveness and cost-utility
(cost per QALY). The latter will be estimated from patients’ responses to the EuroQol
(EQ-5D) questionnaires using the York MVH tariff. Uncertainty regarding specific
parameters within the analysis will be subjected to a sensitivity analysis, and
uncertainty around the point estimate of the cost utility ratio will be represented using
cost-effectiveness acceptability curves. To inform the economic analyses, the
preferences of potential patients and clinicians between carotid endarterectomy and
carotid stenting given various differences in outcomes will be explored using the
technique of conjoint analysis (discrete choice experiments). A sample of members of
the general population (matched to the ICSS patients) and clinicians will be asked to
complete a questionnaire after completion of randomisation, informed by the
preliminary safety results. Preliminary work, structuring and piloting the conjoint
analyses, will be undertaken earlier.
Stenosis after treatment 25.8
Patency of the carotid artery will be monitored by Doppler ultrasound at a minimum of
30 days after treatment and then annually after randomisation. Restenosis should only
be treated by further angioplasty or surgery if the patient has relevant new symptoms.
Restenosis is usually the result of smooth muscle hypertrophy or neo-intimal
hyperplasia, rather than recurrence of atherosclerosis and hence may not cause
embolic stroke. Asymptomatic restenosis will not be an indication to retreat the lesion
because the risk of disabling symptoms after restenosis is not known.
Crossovers 25.9
Crossovers before any attempt to treat the randomised artery by the allocated
treatment will be avoided unless clinically essential, because the trial data will be
analysed by intention to treat. Patients who are randomised to stenting after ultrasound
or other non-invasive investigation, in whom subsequent angiography prior to stenting
reveals one or more exclusion criteria, should be treated by surgery, if appropriate, or
medical care only if surgery is not appropriate (e.g. because the stenosis is less than
50%). These patients will continue follow up in the trial and will be analysed on an
226
intention-to-treat basis. A similar approach should be taken to patients randomised to
surgery in whom contraindications to surgery emerge after randomisation. Patient
refusal of the treatment to which they are randomised can be minimised by careful
consent. Patients requiring re-treatment because of further symptoms should be re-
treated with whichever treatment is most appropriate. This is also the case if the non-
randomised carotid artery requires treatment. Patients in whom an attempt at stenting
fails may proceed to early surgery if appropriate and vice versa.
Data analysis 25.10
The data will be analysed by intention to treat using standard statistical tests by the trial
statistician. The analyses will compare the treatment groups with respect to the length
of time before treatment failure (i.e. occurrence of an outcome event) by means of the
Mantel-Haenszel chi-squared test and Kaplan-Meier survival curves. Secondary
analysis will compare the proportions of outcome events within 30 days of treatment.
All analyses will be adjusted for centre and predetermined risk factors. Subgroup
analyses will examine risk factors for outcome events and will examine the influence of
different devices, surgical techniques and experience within the trial. Results at
probationary centres will be analysed separately. The results of any interim data
analysis will remain confidential to the trial statistician and Data Monitoring Committee
until after completion or early discontinuation of the trial. Investigators and the Steering
Committee will remain blind until such point.
Publication 25.11
Publication of the results of ICSS will be prepared by the Central Office and circulated
to participating centres for comment prior to submission of the manuscript for
publication on behalf of all the ICSS collaborators.
Ethical Committee approval 25.12
Multicentre Research Ethics Committee approval will be sought in the UK. In addition,
individual centres are expected to obtain local ethical committee approval for the study.
Data Monitoring Committee 25.13
The safety aspects of the trial will be overseen by a Data Monitoring Committee
consisting of an independent neurologist, medical statistician, surgeon and
interventionist. The progress of the study will be assessed at regular intervals
227
determined by the Data Monitoring Committee. During the period of intake to the study,
interim analyses of mortality and of any other information that is available on major
endpoints (including serious adverse events believed to be due to treatment) will be
supplied, in strict confidence, to the chairman of the Data Monitoring Committee, along
with any other analyses that the Committee may request. In the light of these analyses,
the Data Monitoring Committee will advise the chairman of the Steering Committee if,
in their view, the randomised comparisons in ICSS have provided both (i) "proof
beyond reasonable doubt" that for all, or for some, specific types of patients, one
particular treatment is clearly indicated or clearly contraindicated in terms of a net
difference in outcome, and (ii) evidence that might reasonably be expected to influence
materially patient management. Appropriate criteria of proof beyond reasonable doubt
cannot be specified precisely, but a difference of at least 3 standard deviations in an
interim analysis of a major endpoint may be needed to justify halting, or modifying, the
study prematurely. This criterion has the practical advantage that the number of interim
analyses is of little importance.
Steering Committee 25.14
A Steering Committee, consisting of individuals participating in and independent of the
trial with experience in stroke medicine, neurology, vascular surgery, vascular
radiology, interventional neuroradiology, health economics, clinical trials and statistics
will oversee the management of the trial.
Trial organisation 25.15
The study will be organised on behalf of the collaborators by the central office, located
at the UCL Institute of Neurology in London. The office will be responsible for protocol
design, data collection and management, and analysis of the results in consultation
with the Steering and Data Monitoring Committees, but will consult with the
collaborators at an annual meeting and at other times as necessary. Communication
with investigators will also take place via a regular newsletter and the trial website.
25.15.1 Payments to centres
While funding is available, the Lead Institution (UCL Institute of Neurology, London) will
pay the participating centres a one-off payment of £100 for each patient randomised
patient for whom correctly filled-out randomisation, technical data and one month
follow-up forms have been received. Participating centres must invoice to the Lead
Institution within 6 months of receipt of this revised protocol and thereafter 6-monthly in
228
arrears. The Lead Centre may vary or terminate such payments in the future in
accordance with budgetary needs and will inform the participating centre of such
changes as they occur.
25.15.2 Indemnity
ICSS is an academic trial performed as a collaborative effort for the benefit of patients,
and is not performed for, or on behalf of, an industry sponsor. The trial compares two
existing forms of treatment currently used in many hospitals. The various devices
approved for use in the trial are not investigational devices and are required by the
protocol to be marketed and already in use in the carotid artery as recognised by the
CE mark. Hence, the trial is not an industry sponsored test of a new treatment with
unknown hazards. The trial protocol anticipates that some patients may be harmed
inadvertently as a result of treatment in the trial. Indeed, the determination of the rate of
these adverse outcome events is a major aim of the trial. However, we believe that the
risks of these adverse events will be outweighed by the benefits of treatment in either
arm of the trial. The trial protocol does not subject patients to hazards that the patient
would not have encountered if they had received the trial treatments outside the
context of the trial in routine practice. Hence, the organisers of the trial cannot take
responsibility for any harm occurring to patients as a result of partaking in the trial.
Individual investigators and hospitals are required to take responsibility for the
occurrence of any adverse events in the same way as they would do if the treatments
were performed outside the trial.
25.15.3 Website
The trial website contains updated information about the trial together with
downloadable copies of the protocol, trial data collection forms, newsletters and contact
information. The names of the collaborating centres will be included on the website.
The website address is http://www.cavatas.com and all the pages are accessible to the
public, patients and collaborators alike without a password. At present, the data
collection forms cannot be completed on line.
229
26. Appendix V – ICSS-MRI Substudy protocol
Symptomatic and asymptomatic ischaemic and haemorrhagic brain injury following
protected and unprotected stenting versus endarterectomy in the International Carotid
Stenting Study – the ICSS-MRI Substudy.
Summary 26.1
26.1.1 Background and aim
Diffusion weighted imaging (DWI) – a modern magnetic resonance imaging (MRI)
technique – may detect ischaemic brain lesions after carotid interventions in patients
who do not experience symptoms. Previous studies showed that DWI lesions are found
more frequently after endarterectomy than after stenting of carotid stenosis, and more
frequently after stenting without the use of cerebral protection devices than after
protected stenting. However, methodological shortcomings of those non-randomised
studies may account for the observed differences. Moreover, it is not clear how
ischaemic lesions on DWI relate to the risk of clinically apparent cerebrovascular
events (stroke or TIA) associated with the intervention. About one in ten strokes
occurring as a complication of carotid interventions is caused by intracerebral
haemorrhage (ICH), but asymptomatic ICH after carotid interventions has never been
assessed. We therefore aim to study the frequency and significance of symptomatic
and asymptomatic ischaemic and haemorrhagic brain injury in protected and
unprotected stenting and endarterectomy in a randomised trial.
26.1.2 Primary objective
To compare the rate of ischaemic brain injury detectable on MRI after treatment
of symptomatic carotid stenosis by stenting or endarterectomy
26.1.3 Secondary objectives
To test for an interaction between the use of cerebral protection devices and
ischaemic brain injury associated with stenting
To compare the rate of haemorrhagic brain injury detectable on MRI after
treatment of symptomatic carotid stenosis by stenting or endarterectomy
To evaluate the usefulness of ischaemic and haemorrhagic brain lesions visible
on MRI as surrogate markers of the procedural risk of carotid interventions
230
Methods 26.2
Multicentre prospective MRI substudy of the International Carotid Stenting Study
(ICSS). Multimodal MRI to detect ischaemic and haemorrhagic brain injury will be
performed at 3 time points in patients with symptomatic carotid stenosis randomised to
stenting or endarterectomy in ICSS: 1-3 days before, 1-3 days after and again 30±3
days after intervention.
Background and aim 26.3
Carotid stenting has emerged as a treatment alternative to endarterectomy in patients
with symptomatic carotid stenosis. Cerebral protection devices are used in stenting
with the aim of reducing the risk of plaque embolisation during the procedure. Recently
completed randomised trials comparing the safety of stenting and endarterectomy
yielded conflicting results [79] [77]. Concern that stenting without cerebral protection
may be associated with an increased risk of stroke led to the abandonment of
unprotected procedures in one trial [63] but in another trial, there was no difference in
the risk of stenting with and without protection [77]. Although clear evidence that
cerebral protection enhances treatment safety is lacking [253], protection devices are
widely used today, significantly contributing to the cost of carotid stenting.
Until clinical trials investigating the benefit of cerebral protection devices can be
realised, surrogate markers of brain injury may provide further insights into the risk of
protected and unprotected stenting in comparison to endarterectomy. Diffusion
weighted imaging (DWI), a modern magnetic resonance imaging (MRI) technique, may
show ischaemic lesions after carotid interventions even in patients who do not
experience symptoms [214]. In previous studies new ischaemic lesions on DWI were
detected more frequently after stenting than after endarterectomy [254] [255] [256]
[257] [258] [259] [260]. DWI lesions were also more frequent after unprotected stenting
than after protected stenting [215] [261]. However, selection bias and the use of
historical controls may be accountable for the observed differences in these non-
randomised comparisons. Moreover, it is not clear how ischaemic lesions on DWI
relate to the risk of clinically apparent cerebrovascular events (stroke or TIA)
associated with the intervention. Larger studies with randomised treatment allocation
are needed to gain further insight into the significance of asymptomatic DWI lesions
and their potential role as surrogate markers of treatment risk.
231
About one in ten strokes occurring as a complication of carotid interventions is caused
by intracerebral haemorrhage (ICH). A special MRI technique (so-called gradient echo
imaging) allows the detection of small intracerebral bleedings, but has never been
applied to detect ICH after carotid interventions. Asymptomatic ICH may be an
expression of subclinical reperfusion damage following carotid revascularisation.
Thus, there is a clear need to study symptomatic and asymptomatic ischaemic and
haemorrhagic brain injury in protected and unprotected stenting and endarterectomy in
a randomised trial.
Objectives 26.4
The primary objective of this substudy is to compare the risk of ischaemic brain injury
assessed on MRI in patients with symptomatic carotid artery stenosis undergoing
stenting in comparison to those undergoing endarterectomy.
Secondary objectives are: to assess the effect of protection devices on the risk of
ischaemic brain injury associated with stenting, to compare the risk of haemorrhagic
brain injury assessed on MRI in stenting compared to endarterectomy, and to gain
further insight into the usefulness of ischaemic and haemorrhagic brain lesions on MRI
as surrogate markers of the risk of carotid interventions.
Study design 26.5
This project is an MRI-based substudy of a multicentre randomised trial known as the
International Carotid Stenting Study [81], which is comparing the risks and benefits of
stenting and endarterectomy of symptomatic carotid stenosis using clinical endpoints,
e.g. stroke and death.
The ICSS-MRI Substudy allows a randomised comparison of the procedural risk of
symptomatic and asymptomatic ischaemic and haemorrhagic brain injury visible on
MRI between stenting and endarterectomy. The use of cerebral protection devices in
patients undergoing stenting is not subject to randomisation in ICSS. However, the
participating centres systematically use either protected or unprotected stenting. The
risk of brain injury associated with either stenting technique can therefore be compared
to a randomised control group of patients undergoing endarterectomy.
Outcome measures and analyses are defined as follows:
232
26.5.1 Primary outcome measure
Rate of symptomatic and asymptomatic ischaemic brain injury detectable on
MRI after endarterectomy and stenting
26.5.2 Secondary analyses
Interaction between the use of protection devices and ischaemic brain injury in
patients undergoing stenting
Rate of symptomatic and asymptomatic haemorrhagic brain injury detectable on
MRI after endarterectomy and stenting
Relation of brain injury on MRI to risk of stroke during procedure and follow-up
26.5.3 Subject selection
Inclusion criteria
Patients are eligible to participate in the ICSS-MRI Substudy if they are enrolled in the
ICSS trial and separately provide written informed consent to undergo three MRI
exams. Detailed inclusion and exclusion criteria for ICSS are provided elsewhere [81].
In short, patients with recently symptomatic ≥50% carotid stenosis who are equally
suitable for endarterectomy and stenting are eligible for enrolment in ICSS.
Exclusion criteria
Patients with contraindications to MRI, e.g. pacemakers, metallic implants, and
claustrophobia, are excluded from the ICSS-MRI Substudy.
26.5.4 MRI protocol
Patients enrolled in the ICSS-MRI Substudy will undergo three MRI investigations, 1-3
days before, 1-3 days after and 30±3 days after the intervention. The following
sequences will be performed in all three investigations:
Diffusion weighted imaging (DWI) to detect acute ischaemic brain injury
associated with the procedure
Gradient echo T2*-weighted sequences to detect haemorrhagic brain injury
associated with the procedure
233
T1-weighted, T2-weighted and fluid-attenuated inversion recovery (FLAIR)
sequences will be used to assess whether acute brain lesions detected on DWI
after the procedure lead to permanent scarring at 1 month
26.5.5 Data acquisition
Baseline data (such as age, gender, medical risk factors, degree of carotid stenosis,
etc.) will be collected as part of ICSS.
Two researchers will independently score the presence, size and location (vascular
territory) of ischaemic and haemorrhagic lesions on the MRI scans. A third researcher
will review the scans in case of disagreement. The scans will be reported and scored
blind to patient identifiers, treatment, date and time of the scans.
Patients will be clinically examined by a neurologist at the time of MRI examination and
will be followed up after treatment as part of ICSS to determine outcome events
including transient ischaemic attack, stroke, myocardial infarction and death.
26.5.6 Statistical considerations
Statistical analysis
The rates of ischaemic and haemorrhagic brain lesions will be compared between
patients undergoing endarterectomy and stenting using chi-squared and Fisher’s exact
tests. Significance will be declared at p<0.05.
Sample size calculation
Power calculations are based on the primary outcome measure. The two largest series
reported new ischaemic lesions on DWI after carotid endarterectomy in 17% and 34%
of patients respectively [262]. If a rate of new DWI lesions after endarterectomy of 25%
is assumed, a total sample size of 200 patients would have a 90% power to detect a
twofold increase in the DWI lesion rate associated with carotid stenting. Testing the
interaction between the use of cerebral protection devices and the rate of DWI lesions
after stenting would have less power. 126 patients have been enrolled in the ICSS-MRI
Substudy from its initiation in 2004 until July 2007 in two participating centres (Utrecht
and Basel). The projected number of enrolled patients at the end of the ICSS
randomisation period in those centres is 170. The target population of 200 could be
reached with a contribution of 30 patients enrolled in ICSS centres in the UK.
234
26.5.7 Withdrawal of consent
Subjects enrolled in the ICSS-MRI Substudy can withdraw their consent at any time
during the substudy. This will not affect their enrolment in ICSS or the standard of care
they receive.